Patent Application
20220114251
Embodiments described herein generally relate to network monitoring and tuning and in particular, to a system and method for implementing reputation management and intent-based security mechanisms.
Edge computing, at a general level, refers to the transition of compute and storage resources closer to endpoint devices (e.g., consumer computing devices, user equipment, etc.) in order to optimize total cost of ownership, reduce application latency, improve service capabilities, and improve compliance with security or data privacy requirements. Edge computing may, in some scenarios, provide a cloud-like distributed service that offers orchestration and management for applications among many types of storage and compute resources. As a result, some implementations of edge computing have been referred to as the “edge cloud” or the “fog”, as powerful computing resources previously available only in large remote data centers are moved closer to endpoints and made available for use by consumers at the “edge” of the network.
Edge computing use cases in mobile network settings have been developed for integration with multi-access edge computing (MEC) approaches, also known as “mobile edge computing.” MEC approaches are designed to allow application developers and content providers to access computing capabilities and an information technology (IT) service environment in dynamic mobile network settings at the edge of the network. Limited standards have been developed by the European Telecommunications Standards Institute (ETSI) industry specification group (ISG) to define common interfaces for operation of MEC systems, platforms, hosts, services, and applications.
Edge computing, MEC, and related technologies attempt to provide reduced latency, increased responsiveness, and more available computing power than offered in traditional cloud network services and wide area network connections. However, the integration of mobility and dynamically launched services to some mobile use and device processing use cases has led to limitations and concerns with orchestration, functional coordination, and resource management, especially in complex mobility settings where many participants (devices, hosts, tenants, service providers, operators) are involved.
In a similar manner, Internet of Things (IoT) networks and devices are designed to offer a distributed compute arrangement, from a variety of endpoints. IoT devices are physical or virtualized objects that may communicate on a network, and may include sensors, actuators, and other input/output components, which may be used to collect data or perform actions in a real-world environment. For example, IoT devices may include low-powered endpoint devices that are embedded or attached to everyday things, such as buildings, vehicles, packages, etc., to provide an additional level of artificial sensory perception of those things. Recently, IoT devices have become more popular and thus applications using these devices have proliferated.
The deployment of various Edge, Fog, MEC, and IoT networks, devices, and services have introduced several advanced use cases and scenarios occurring at and towards the edge of the network. However, these advanced use cases have also introduced several corresponding technical challenges relating to orchestration, security, processing and network resources, service availability and efficiency, ensuring quality of service, among many other issues, especially as more types of computing systems and configurations are deployed.
In the drawings, which are not necessarily drawn to scale, like numerals may describe similar components in different views. Like numerals having different letter suffixes may represent different instances of similar components. Some embodiments are illustrated by way of example, and not limitation, in the figures of the accompanying drawings in which:
In the drawings, which are not necessarily drawn to scale, like numerals may describe similar components in different views. Like numerals having different letter suffixes may represent different instances of similar components. Some embodiments are illustrated by way of example, and not limitation, in the figures of the accompanying drawings in which:
Systems and methods described herein provide for reputation management and intent-based security mechanisms. Current orchestration solutions use a very imperative way to achieving quality of service (QoS) management. QoS management is mostly built around requesting the correct quantity of resources (e.g., number of virtual central processing units (vCPUs)), or hardware features to support the workloads at the cloud and edge. Performing orchestration this way has several issues. It creates unwanted vendor-lock in, as specific feature sets specified by one vendor may not be available by equipment provided by another vendor. Further, experience shows that incorrect information is declared in the QoS request and hence leads to sub-optimal performance Finally, these declarations have context that are not included in the QoS management scheme. For instance, the type of platform the application is running on (e.g., Xeon type core vs Atom-based core processors) may affect the provisioning decisions to satisfy a certain QoS request.
Furthermore, as the workload is moving from a monolithic style to a microservices style, it becomes harder for customers to select the correct resource allocations for the workloads by themselves. This may lead to the customers overprovisioning the resources, which results in higher cost.
It is important that as hardware platforms become more heterogenous (e.g., using multiple XPUs instead of a single CPU), the user should be abstracted from needing to define the multitude of contextual details in an orchestration request, and instead be able to focus on what it wants to achieve a certain set of objectives for the services at play. An intent-driven model provides this abstraction for the user and results in good performance for the service owner as well as good return on investment for the resource owner. Thus, what is needed is a way to map intents (defined as service level objectives) across a set of systems and their resources to achieve required levels of quality of service.
A service level agreement (SLA) is an agreement between a service provider and a client. SLAs are contracts between third parties that outline service targets or other business objectives. SLAs may outline penalties for parties who fail to satisfy a target. For example, a partial refund of fees may be enforced when a service is unavailable for more than some threshold percentage over a 90 day period.
Service level objectives (SLOs) provide precise numerical targets for various system capabilities. Typical SLOs are oriented around service availability or uptime, service latency, service bandwidth provisioning, etc. SLOs are often expressed as percentages, such as requiring “99.9% uptime over a 30-day period,” or “Response time of less than 100 ms on at least 95% of the requests received.”
Key performance indicators (KPIs) are quantifiable measures of performance over time for a specific objective. KPI metrics are collected to measure SLO policies. Example KPIs include, but are not limited to a number of service errors, downtime, cache misses, frames per second (FPS), latency, instructions per cycle (IPC), security faults, failed logins, etc. KPIs may be derived values (e.g., average latency over a moving 30 second window). KPIs may also be based on comparisons (e.g., percent different from historical average).
Service quality objectives may be captured as statistical and deterministic specifications that form a layered, top-down structure, but not constrained to be in a tree or DAG pattern. For example, a meta-language may be used to express key performance indicators (KPIs) and how they are combined to evaluate an overall quality of service (QoS) or Service Level Objective (SLO). Various ranges, comparisons, thresholds, constraints, or the like, on the KPIs may be expressed using a meta-language. Most real-world software solutions have to satisfy many intricate demands for performance, scale, accuracy and availability, and these demands are not necessarily fixed for all time or for all situations. Just as programs in a language like C or Python permit a programmer a high degree of flexibility in expressing what is to be computed, a meta-language may be used that expresses the dynamic requirements that need to be met by schedulers of resources. A meta-program written in such a meta-language may introduce variables that, for example, reflect various calibrations that are performed on the behaviors of programs, runtimes, operating systems, infrastructural services, networking gear, and so on. These variables are then woven into higher level variables and results of the meta-programs, and in turn drive corrective, reactive, alertive, and other actions from meta-programs. The structuring is layered so that meta-programs are designed for abstraction and reduction of detail and for convergence to the specified objectives at higher levels.
An open architecture is provided so that the layering of policies, heuristics, etc., is not confined by rules of logic alone, but permits the ability to plug-in data-programmed intelligences such as neural networks, support vector machines (SVMs), decision trees, etc. The overall goal is to redirect a portion of computing to such meta-programs and avoid making a strict division between the computations that are part of an application's structure and the computations that are designed to meet a service quality objective during the execution of the application. Thus, the ordinary distinction between a control/administrative plane like Kubernetes and a containerized application under which it runs is blurred, and information about the intent and the performance to that intent is allowed to flow bidirectionally at different layers of depth. Accordingly, auto-corrective and auto-indicative capabilities can be co-engineered between the flows in service-objectives and service-quality-objectives. A benefit is that software begins to achieve self-adaptive or co-adaptive accommodation with the environments in which it is activated.
Statistical admission control policies that attempt to maximize Service Level Agreements (SLA) satisfaction are available and are related to jitter-sensitive network systems. However, as abstraction moves upward and scheduling in serverless systems or highly virtualized is implemented and just-in-time assignments of resources through containers created on demand, the complexity of scheduling resources continues to increase. Very large data centers can provision significant amounts of capacity to perform reactive load-balancing and can defer admission denials to all but the most extreme bursts of demand, but smaller, and less elastic providers of infrastructure have to make tough calls even under statistical policies because when a cutoff is for an outlier tail latency. It is extremely difficult to honor large bursts in demand that can cause a cascade of latency pushouts. As an alternative, the systems and methods described herein implement SLAs that are both nested and graduated and are thus burst-accommodative.
Imagine a token-bucket model with a single queue at a rate limited server: a statistical projection for the newest arrival has a mean response time that is proportional to the queue length seen by the new arrival. A naïve policy would be to reject the new arrival if this projected response time violates a tail latency constraint; a slightly less naïve policy would be to reject the new arrival probabilistically, depending on the queue length and the number of violations that have already occurred. Now imagine a generalization in which the baseline single queue, single rate server is split into a two-level policy in which the newest arrival is placed in the second queue where it is guaranteed a lower service rate but permitted a more relaxed tail latency. Thus, for example, if the first queue had a P99 latency of 10 ms, the second queue may have a combination SLO with a P95 latency of 10 ms and a P99 latency of 20 ms. Because extreme bursts are rare, the fractional capacity devoted to the second queue need not be high, because its average queue length is small. In a data center, the movement to a second queue is effectively reassigning the marginal request (the one that is likely to violate a stringent latency bound), to a less utilized but less resourced overflow cluster with correspondingly more relaxed SLO. This scheme can be generalized so that as the second queue approaches its saturation for response time, a third queue with even more liberal SLO absorbs the overflow. In this way, a compound nested SLOs within an SLA can be crafted that is adaptive to bursts. The provider can negotiate a similarly graduated cost model in which the savings from not having to provision a lot of peak capacity to meet an inelastic SLA are passed back to customers who accept a nested SLA.
As opposed to Hierarchical Service Level Agreements, the systems and methods described here introduce SLAs that are both nested and graduated. Instead of relying on classic thresholding, and instead of having hard set “single” clause rules to check, a nested sub-clause of the SLA can be evaluated in combination with other sub-clauses to evaluate the overall SLA. This allows more complex rules. Also, “parsing” of the sub-clauses allows results of each clause to create sharing of unused resources. This approach allows more flexibility in the SLA rules and better use of cluster resources. The flexibility can be introduced as an “intent” not a resource specification, which then is mapped to nested/graduated SLA rules which are then monitored and enforced.
In production deployments a common pattern is to deploy new instances of an application on a portion of the environment and test with a small subset of the user base before rolling out to the wider population (i.e., Canary roll-out). Acknowledging that the SLA mapping to lower and lower levels of Service Level Objectives requires an iterative approach (for both optimization and remediation purposes), the new workflow for this SLA decomposition also includes the potential to partially deploy in either a limited end-to-end fashion or as sub-components of the E2E solution to determine impact on SLA adherence.
In summary, the systems and methods described herein provide a new way to achieve orchestration, moving from a current model to an objective driven approach, in which the customer expresses only the intent (e.g., latency, throughput, and reliability properties) and the orchestration stack itself sets up the platform to achieve that intent. As serverless infrastructure becomes more common, cloud-native (microservice) style workloads will eventually demand a way of intent driven orchestration. These functions and others are described in more detail below.
Compute, memory, and storage are scarce resources, and generally decrease depending on the Edge location (e.g., fewer processing resources being available at consumer endpoint devices, than at a base station, than at a central office). However, the closer that the Edge location is to the endpoint (e.g., user equipment (UE)), the more that space and power is often constrained. Thus, Edge computing attempts to reduce the amount of resources needed for network services, through the distribution of more resources which are located closer both geographically and in network access time. In this manner, Edge computing attempts to bring the compute resources to the workload data where appropriate, or, bring the workload data to the compute resources.
The following describes aspects of an Edge cloud architecture that covers multiple potential deployments and addresses restrictions that some network operators or service providers may have in their own infrastructures. These include, variation of configurations based on the Edge location (because edges at a base station level, for instance, may have more constrained performance and capabilities in a multi-tenant scenario); configurations based on the type of compute, memory, storage, fabric, acceleration, or like resources available to Edge locations, tiers of locations, or groups of locations; the service, security, and management and orchestration capabilities; and related objectives to achieve usability and performance of end services. These deployments may accomplish processing in network layers that may be considered as “near Edge”, “close Edge”, “local Edge”, “middle Edge”, or “far Edge” layers, depending on latency, distance, and timing characteristics.
Edge computing is a developing paradigm where computing is performed at or closer to the “Edge” of a network, typically through the use of a compute platform (e.g., x86 or ARM compute hardware architecture) implemented at base stations, gateways, network routers, or other devices which are much closer to endpoint devices producing and consuming the data. For example, Edge gateway servers may be equipped with pools of memory and storage resources to perform computation in real-time for low latency use-cases (e.g., autonomous driving or video surveillance) for connected client devices. Or as an example, base stations may be augmented with compute and acceleration resources to directly process service workloads for connected user equipment, without further communicating data via backhaul networks. Or as another example, central office network management hardware may be replaced with standardized compute hardware that performs virtualized network functions and offers compute resources for the execution of services and consumer functions for connected devices. Within Edge computing networks, there may be scenarios in services which the compute resource will be “moved” to the data, as well as scenarios in which the data will be “moved” to the compute resource. Or as an example, base station compute, acceleration and network resources can provide services in order to scale to workload demands on an as needed basis by activating dormant capacity (subscription, capacity on demand) in order to manage corner cases, emergencies or to provide longevity for deployed resources over a significantly longer implemented lifecycle.
Examples of latency, resulting from network communication distance and processing time constraints, may range from less than a millisecond (ms) when among the endpoint layer 200, under 5 ms at the Edge devices layer 210, to even between 10 to 40 ms when communicating with nodes at the network access layer 220. Beyond the Edge cloud 110 are core network 230 and cloud data center 240 layers, each with increasing latency (e.g., between 50-60 ms at the core network layer 230, to 100 or more ms at the cloud data center layer). As a result, operations at a core network data center 235 or a cloud data center 245, with latencies of at least 50 to 100 ms or more, will not be able to accomplish many time-critical functions of the use cases 205. Each of these latency values are provided for purposes of illustration and contrast; it will be understood that the use of other access network mediums and technologies may further reduce the latencies. In some examples, respective portions of the network may be categorized as “close Edge”, “local Edge”, “near Edge”, “middle Edge”, or “far Edge” layers, relative to a network source and destination. For instance, from the perspective of the core network data center 235 or a cloud data center 245, a central office or content data network may be considered as being located within a “near Edge” layer (“near” to the cloud, having high latency values when communicating with the devices and endpoints of the use cases 205), whereas an access point, base station, on-premise server, or network gateway may be considered as located within a “far Edge” layer (“far” from the cloud, having low latency values when communicating with the devices and endpoints of the use cases 205). It will be understood that other categorizations of a particular network layer as constituting a “close”, “local”, “near”, “middle”, or “far” Edge may be based on latency, distance, number of network hops, or other measurable characteristics, as measured from a source in any of the network layers 200-240.
The various use cases 205 may access resources under usage pressure from incoming streams, due to multiple services utilizing the Edge cloud. To achieve results with low latency, the services executed within the Edge cloud 110 balance varying requirements in terms of: (a) Priority (throughput or latency) and Quality of Service (QoS) (e.g., traffic for an autonomous car may have higher priority than a temperature sensor in terms of response time requirement; or, a performance sensitivity/bottleneck may exist at a compute/accelerator, memory, storage, or network resource, depending on the application); (b) Reliability and Resiliency (e.g., some input streams need to be acted upon and the traffic routed with mission-critical reliability, where as some other input streams may be tolerate an occasional failure, depending on the application); and (c) Physical constraints (e.g., power, cooling and form-factor).
The end-to-end service view for these use cases involves the concept of a service-flow and is associated with a transaction. The transaction details the overall service requirement for the entity consuming the service, as well as the associated services for the resources, workloads, workflows, and business functional and business level requirements. The services executed with the “terms” described may be managed at each layer in a way to assure real time, and runtime contractual compliance for the transaction during the lifecycle of the service. When a component in the transaction is missing its agreed to SLA, the system as a whole (components in the transaction) may provide the ability to (1) understand the impact of the SLA violation, and (2) augment other components in the system to resume overall transaction SLA, and (3) implement steps to remediate.
Thus, with these variations and service features in mind, Edge computing within the Edge cloud 110 may provide the ability to serve and respond to multiple applications of the use cases 205 (e.g., object tracking, video surveillance, connected cars, etc.) in real-time or near real-time, and meet ultra-low latency requirements for these multiple applications. These advantages enable a whole new class of applications (Virtual Network Functions (VNFs), Function as a Service (FaaS), Edge as a Service (EaaS), standard processes, etc.), which cannot leverage cloud computing due to latency or other limitations.
However, with the advantages of Edge computing comes the following caveats. The devices located at the Edge are often resource constrained and therefore there is pressure on usage of Edge resources. Typically, this is addressed through the pooling of memory and storage resources for use by multiple users (tenants) and devices. The Edge may be power and cooling constrained and therefore the power usage needs to be accounted for by the applications that are consuming the most power. There may be inherent power-performance tradeoffs in these pooled memory resources, as many of them are likely to use emerging memory technologies, where more power requires greater memory bandwidth. Likewise, improved security of hardware and root of trust trusted functions are also required, because Edge locations may be unmanned and may even need permissioned access (e.g., when housed in a third-party location). Such issues are magnified in the Edge cloud 110 in a multi-tenant, multi-owner, or multi-access setting, where services and applications are requested by many users, especially as network usage dynamically fluctuates and the composition of the multiple stakeholders, use cases, and services changes.
At a more generic level, an Edge computing system may be described to encompass any number of deployments at the previously discussed layers operating in the Edge cloud 110 (network layers 200-240), which provide coordination from client and distributed computing devices. One or more Edge gateway nodes, one or more Edge aggregation nodes, and one or more core data centers may be distributed across layers of the network to provide an implementation of the Edge computing system by or on behalf of a telecommunication service provider (“telco”, or “TSP”), internet-of-things service provider, cloud service provider (CSP), enterprise entity, or any other number of entities. Various implementations and configurations of the Edge computing system may be provided dynamically, such as when orchestrated to meet service objectives.
Consistent with the examples provided herein, a client compute node may be embodied as any type of endpoint component, device, appliance, or other thing capable of communicating as a producer or consumer of data. Further, the label “node” or “device” as used in the Edge computing system does not necessarily mean that such node or device operates in a client or agent/minion/follower role; rather, any of the nodes or devices in the Edge computing system refer to individual entities, nodes, or subsystems which include discrete or connected hardware or software configurations to facilitate or use the Edge cloud 110.
As such, the Edge cloud 110 is formed from network components and functional features operated by and within Edge gateway nodes, Edge aggregation nodes, or other Edge compute nodes among network layers 210-230. The Edge cloud 110 thus may be embodied as any type of network that provides Edge computing and/or storage resources which are proximately located to radio access network (RAN) capable endpoint devices (e.g., mobile computing devices, IoT devices, smart devices, etc.), which are discussed herein. In other words, the Edge cloud 110 may be envisioned as an “Edge” which connects the endpoint devices and traditional network access points that serve as an ingress point into service provider core networks, including mobile carrier networks (e.g., Global System for Mobile Communications (GSM) networks, Long-Term Evolution (LTE) networks, 5G/6G networks, etc.), while also providing storage and/or compute capabilities. Other types and forms of network access (e.g., Wi-Fi, long-range wireless, wired networks including optical networks) may also be utilized in place of or in combination with such 3GPP carrier networks.
The network components of the Edge cloud 110 may be servers, multi-tenant servers, appliance computing devices, and/or any other type of computing devices. For example, the Edge cloud 110 may include an appliance computing device that is a self-contained electronic device including a housing, a chassis, a case or a shell. In some circumstances, the housing may be dimensioned for portability such that it can be carried by a human and/or shipped. Example housings may include materials that form one or more exterior surfaces that partially or fully protect contents of the appliance, in which protection may include weather protection, hazardous environment protection (e.g., EMI, vibration, extreme temperatures), and/or enable submergibility. Example housings may include power circuitry to provide power for stationary and/or portable implementations, such as AC power inputs, DC power inputs, AC/DC or DC/AC converter(s), power regulators, transformers, charging circuitry, batteries, wired inputs and/or wireless power inputs. Example housings and/or surfaces thereof may include or connect to mounting hardware to enable attachment to structures such as buildings, telecommunication structures (e.g., poles, antenna structures, etc.) and/or racks (e.g., server racks, blade mounts, etc.). Example housings and/or surfaces thereof may support one or more sensors (e.g., temperature sensors, vibration sensors, light sensors, acoustic sensors, capacitive sensors, proximity sensors, etc.). One or more such sensors may be contained in, carried by, or otherwise embedded in the surface and/or mounted to the surface of the appliance. Example housings and/or surfaces thereof may support mechanical connectivity, such as propulsion hardware (e.g., wheels, propellers, etc.) and/or articulating hardware (e.g., robot arms, pivotable appendages, etc.). In some circumstances, the sensors may include any type of input devices such as user interface hardware (e.g., buttons, switches, dials, sliders, etc.). In some circumstances, example housings include output devices contained in, carried by, embedded therein and/or attached thereto. Output devices may include displays, touchscreens, lights, LEDs, speakers, I/O ports (e.g., USB), etc. In some circumstances, Edge devices are devices presented in the network for a specific purpose (e.g., a traffic light), but may have processing and/or other capacities that may be utilized for other purposes. Such Edge devices may be independent from other networked devices and may be provided with a housing having a form factor suitable for its primary purpose; yet be available for other compute tasks that do not interfere with its primary task. Edge devices include Internet of Things devices. The appliance computing device may include hardware and software components to manage local issues such as device temperature, vibration, resource utilization, updates, power issues, physical and network security, etc. Example hardware for implementing an appliance computing device is described in conjunction with
In
In the example of
It should be understood that some of the devices in 410 are multi-tenant devices where Tenant 1 may function within a tenant1 ‘slice’ while a Tenant 2 may function within a tenant2 slice (and, in further examples, additional or sub-tenants may exist; and each tenant may even be specifically entitled and transactionally tied to a specific set of features all the way day to specific hardware features). A trusted multi-tenant device may further contain a tenant specific cryptographic key such that the combination of key and slice may be considered a “root of trust” (RoT) or tenant specific RoT. A RoT may further be computed dynamically composed using a DICE (Device Identity Composition Engine) architecture such that a single DICE hardware building block may be used to construct layered trusted computing base contexts for layering of device capabilities (such as a Field Programmable Gate Array (FPGA)). The RoT may further be used for a trusted computing context to enable a “fan-out” that is useful for supporting multi-tenancy. Within a multi-tenant environment, the respective Edge nodes 422, 424 may operate as security feature enforcement points for local resources allocated to multiple tenants per node. Additionally, tenant runtime and application execution (e.g., in instances 432, 434) may serve as an enforcement point for a security feature that creates a virtual Edge abstraction of resources spanning potentially multiple physical hosting platforms. Finally, the orchestration functions 460 at an orchestration entity may operate as a security feature enforcement point for marshalling resources along tenant boundaries.
Edge computing nodes may partition resources (memory, central processing unit (CPU), graphics processing unit (GPU), interrupt controller, input/output (I/O) controller, memory controller, bus controller, etc.) where respective partitionings may contain a RoT capability and where fan-out and layering according to a DICE model may further be applied to Edge Nodes. Cloud computing nodes often use containers, FaaS engines, Servlets, servers, or other computation abstraction that may be partitioned according to a DICE layering and fan-out structure to support a RoT context for each. Accordingly, the respective RoTs spanning devices 410, 422, and 440 may coordinate the establishment of a distributed trusted computing base (DTCB) such that a tenant-specific virtual trusted secure channel linking all elements end to end can be established.
Further, it will be understood that a container may have data or workload specific keys protecting its content from a previous Edge node. As part of migration of a container, a pod controller at a source Edge node may obtain a migration key from a target Edge node pod controller where the migration key is used to wrap the container-specific keys. When the container/pod is migrated to the target Edge node, the unwrapping key is exposed to the pod controller that then decrypts the wrapped keys. The keys may now be used to perform operations on container specific data. The migration functions may be gated by properly attested Edge nodes and pod managers (as described above).
In further examples, an Edge computing system is extended to provide for orchestration of multiple applications through the use of containers (a contained, deployable unit of software that provides code and needed dependencies) in a multi-owner, multi-tenant environment. A multi-tenant orchestrator may be used to perform key management, trust anchor management, and other security functions related to the provisioning and lifecycle of the trusted ‘slice’ concept in
For instance, each Edge node 422, 424 may implement the use of containers, such as with the use of a container “pod” 426, 428 providing a group of one or more containers. In a setting that uses one or more container pods, a pod controller or orchestrator is responsible for local control and orchestration of the containers in the pod. Various Edge node resources (e.g., storage, compute, services, depicted with hexagons) provided for the respective Edge slices 432, 434 are partitioned according to the needs of each container.
With the use of container pods, a pod controller oversees the partitioning and allocation of containers and resources. The pod controller receives instructions from an orchestrator (e.g., orchestrator 460) that instructs the controller on how best to partition physical resources and for what duration, such as by receiving key performance indicator (KPI) targets based on SLA contracts. The pod controller determines which container requires which resources and for how long in order to complete the workload and satisfy the SLA. The pod controller also manages container lifecycle operations such as: creating the container, provisioning it with resources and applications, coordinating intermediate results between multiple containers working on a distributed application together, dismantling containers when workload completes, and the like. Additionally, a pod controller may serve a security role that prevents assignment of resources until the right tenant authenticates or prevents provisioning of data or a workload to a container until an attestation result is satisfied.
Also, with the use of container pods, tenant boundaries can still exist but in the context of each pod of containers. If each tenant specific pod has a tenant specific pod controller, there will be a shared pod controller that consolidates resource allocation requests to avoid typical resource starvation situations. Further controls may be provided to ensure attestation and trustworthiness of the pod and pod controller. For instance, the orchestrator 460 may provision an attestation verification policy to local pod controllers that perform attestation verification. If an attestation satisfies a policy for a first tenant pod controller but not a second tenant pod controller, then the second pod could be migrated to a different Edge node that does satisfy it. Alternatively, the first pod may be allowed to execute and a different shared pod controller is installed and invoked prior to the second pod executing.
The system arrangements of depicted in
In the context of
In further examples, aspects of software-defined or controlled silicon hardware, and other configurable hardware, may integrate with the applications, functions, and services an Edge computing system. Software defined silicon (SDSi) may be used to ensure the ability for some resource or hardware ingredient to fulfill a contract or service level agreement, based on the ingredient's ability to remediate a portion of itself or the workload (e.g., by an upgrade, reconfiguration, or provision of new features within the hardware configuration itself).
In further examples, any of the compute nodes or devices discussed with reference to the present Edge computing systems and environment may be fulfilled based on the components depicted in
In the simplified example depicted in
The compute node 600 may be embodied as any type of engine, device, or collection of devices capable of performing various compute functions. In some examples, the compute node 600 may be embodied as a single device such as an integrated circuit, an embedded system, a field-programmable gate array (FPGA), a system-on-a-chip (SOC), or other integrated system or device. In the illustrative example, the compute node 600 includes or is embodied as a processor (also referred to herein as “processor circuitry”) 604 and a memory (also referred to herein as “memory circuitry”) 606. The processor 604 may be embodied as any type of processor(s) capable of performing the functions described herein (e.g., executing an application). For example, the processor 604 may be embodied as a multi-core processor(s), a microcontroller, a processing unit, a specialized or special purpose processing unit, or other processor or processing/controlling circuit.
In some examples, the processor 604 may be embodied as, include, or be coupled to an FPGA, an application specific integrated circuit (ASIC), reconfigurable hardware or hardware circuitry, or other specialized hardware to facilitate performance of the functions described herein. Also in some examples, the processor 604 may be embodied as a specialized x-processing unit (x-PU) also known as a data processing unit (DPU), infrastructure processing unit (IPU), or network processing unit (NPU). Such an xPU may be embodied as a standalone circuit or circuit package, integrated within an SOC, or integrated with networking circuitry (e.g., in a SmartNIC, or enhanced SmartNIC), acceleration circuitry, storage devices, storage disks, or AI hardware (e.g., GPUs or programmed FPGAs). Such an x-PU may be designed to receive, retrieve and/or otherwise obtain programming to process one or more data streams and perform specific tasks and actions for the data streams (such as hosting microservices, performing service management or orchestration, organizing or managing server or data center hardware, managing service meshes, or collecting and distributing telemetry), outside of the CPU or general purpose processing hardware. However, it will be understood that an x-PU, a SOC, a CPU, and other variations of the processor 604 may work in coordination with each other to execute many types of operations and instructions within and on behalf of the compute node 600.
The memory 606 may be embodied as any type of volatile (e.g., dynamic random access memory (DRAM), etc.) or non-volatile memory or data storage capable of performing the functions described herein. Volatile memory may be a storage medium that requires power to maintain the state of data stored by the medium. Non-limiting examples of volatile memory may include various types of random access memory (RAM), such as DRAM or static random access memory (SRAM). One particular type of DRAM that may be used in a memory module is synchronous dynamic random access memory (SDRAM).
In an example, the memory device (e.g., memory circuitry) is any number of block addressable memory devices, such as those based on NAND or NOR technologies (for example, Single-Level Cell (“SLC”), Multi-Level Cell (“MLC”), Quad-Level Cell (“QLC”), Tri-Level Cell (“TLC”), or some other NAND). In some examples, the memory device(s) includes a byte-addressable write-in-place three dimensional crosspoint memory device, or other byte addressable write-in-place non-volatile memory (NVM) devices, such as single or multi-level Phase Change Memory (PCM) or phase change memory with a switch (PCMS), NVM devices that use chalcogenide phase change material (for example, chalcogenide glass), resistive memory including metal oxide base, oxygen vacancy base and Conductive Bridge Random Access Memory (CB-RAM), nanowire memory, ferroelectric transistor random access memory (FeTRAM), magneto resistive random access memory (MRAM) that incorporates memristor technology, spin transfer torque (STT)-MRAM, a spintronic magnetic junction memory based device, a magnetic tunneling junction (MTJ) based device, a DW (Domain Wall) and SOT (Spin Orbit Transfer) based device, a thyristor based memory device, a combination of any of the above, or other suitable memory. A memory device may also include a three-dimensional crosspoint memory device (e.g., Intel® 3D XPoint™ memory), or other byte addressable write-in-place nonvolatile memory devices. The memory device may refer to the die itself and/or to a packaged memory product. In some examples, 3D crosspoint memory (e.g., Intel® 3D XPoint™ memory) may include a transistor-less stackable cross point architecture in which memory cells sit at the intersection of word lines and bit lines and are individually addressable and in which bit storage is based on a change in bulk resistance. In some examples, all or a portion of the memory 606 may be integrated into the processor 604. The memory 606 may store various software and data used during operation such as one or more applications, data operated on by the application(s), libraries, and drivers.
In some examples, resistor-based and/or transistor-less memory architectures include nanometer scale phase-change memory (PCM) devices in which a volume of phase-change material resides between at least two electrodes. Portions of the example phase-change material exhibit varying degrees of crystalline phases and amorphous phases, in which varying degrees of resistance between the at least two electrodes can be measured. In some examples, the phase-change material is a chalcogenide-based glass material. Such resistive memory devices are sometimes referred to as memristive devices that remember the history of the current that previously flowed through them. Stored data is retrieved from example PCM devices by measuring the electrical resistance, in which the crystalline phases exhibit a relatively lower resistance value(s) (e.g., logical “0”) when compared to the amorphous phases having a relatively higher resistance value(s) (e.g., logical “1”).
Example PCM devices store data for long periods of time (e.g., approximately 10 years at room temperature). Write operations to example PCM devices (e.g., set to logical “0”, set to logical “1”, set to an intermediary resistance value) are accomplished by applying one or more current pulses to the at least two electrodes, in which the pulses have a particular current magnitude and duration. For instance, a long low current pulse (SET) applied to the at least two electrodes causes the example PCM device to reside in a low-resistance crystalline state, while a comparatively short high current pulse (RESET) applied to the at least two electrodes causes the example PCM device to reside in a high-resistance amorphous state.
In some examples, implementation of PCM devices facilitates non-von Neumann computing architectures that enable in-memory computing capabilities. Generally speaking, traditional computing architectures include a central processing unit (CPU) communicatively connected to one or more memory devices via a bus. As such, a finite amount of energy and time is consumed to transfer data between the CPU and memory, which is a known bottleneck of von Neumann computing architectures. However, PCM devices minimize and, in some cases, eliminate data transfers between the CPU and memory by performing some computing operations in-memory. Stated differently, PCM devices both store information and execute computational tasks. Such non-von Neumann computing architectures may implement vectors having a relatively high dimensionality to facilitate hyperdimensional computing, such as vectors having 10,000 bits. Relatively large bit width vectors enable computing paradigms modeled after the human brain, which also processes information analogous to wide bit vectors.
The compute circuitry 602 is communicatively coupled to other components of the compute node 600 via the I/O subsystem 608, which may be embodied as circuitry and/or components to facilitate input/output operations with the compute circuitry 602 (e.g., with the processor 604 and/or the main memory 606) and other components of the compute circuitry 602. For example, the I/O subsystem 608 may be embodied as, or otherwise include, memory controller hubs, input/output control hubs, integrated sensor hubs, firmware devices, communication links (e.g., point-to-point links, bus links, wires, cables, light guides, printed circuit board traces, etc.), and/or other components and subsystems to facilitate the input/output operations. In some examples, the I/O subsystem 608 may form a portion of a system-on-a-chip (SoC) and be incorporated, along with one or more of the processor 604, the memory 606, and other components of the compute circuitry 602, into the compute circuitry 602.
The one or more illustrative data storage devices/disks 610 may be embodied as one or more of any type(s) of physical device(s) configured for short-term or long-term storage of data such as, for example, memory devices, memory, circuitry, memory cards, flash memory, hard disk drives, solid-state drives (SSDs), and/or other data storage devices/disks. Individual data storage devices/disks 610 may include a system partition that stores data and firmware code for the data storage device/disk 610. Individual data storage devices/disks 610 may also include one or more operating system partitions that store data files and executables for operating systems depending on, for example, the type of compute node 600.
The communication circuitry 612 may be embodied as any communication circuit, device, or collection thereof, capable of enabling communications over a network between the compute circuitry 602 and another compute device (e.g., an Edge gateway of an implementing Edge computing system). The communication circuitry 612 may be configured to use any one or more communication technology (e.g., wired or wireless communications) and associated protocols (e.g., a cellular networking protocol such a 3GPP 4G or 5G standard, a wireless local area network protocol such as IEEE 802.11/Wi-Fi®, a wireless wide area network protocol, Ethernet, Bluetooth®, Bluetooth Low Energy, a IoT protocol such as IEEE 802.15.4 or ZigBee®, low-power wide-area network (LPWAN) or low-power wide-area (LPWA) protocols, etc.) to effect such communication.
The illustrative communication circuitry 612 includes a network interface controller (NIC) 620, which may also be referred to as a host fabric interface (HFI). The NIC 620 may be embodied as one or more add-in-boards, daughter cards, network interface cards, controller chips, chipsets, or other devices that may be used by the compute node 600 to connect with another compute device (e.g., an Edge gateway node). In some examples, the NIC 620 may be embodied as part of a system-on-a-chip (SoC) that includes one or more processors, or included on a multichip package that also contains one or more processors. In some examples, the NIC 620 may include a local processor (not shown) and/or a local memory (not shown) that are both local to the NIC 620. In such examples, the local processor of the NIC 620 may be capable of performing one or more of the functions of the compute circuitry 602 described herein. Additionally, or alternatively, in such examples, the local memory of the NIC 620 may be integrated into one or more components of the client compute node at the board level, socket level, chip level, and/or other levels.
Additionally, in some examples, a respective compute node 600 may include one or more peripheral devices 614. Such peripheral devices 614 may include any type of peripheral device found in a compute device or server such as audio input devices, a display, other input/output devices, interface devices, and/or other peripheral devices, depending on the particular type of the compute node 600. In further examples, the compute node 600 may be embodied by a respective Edge compute node (whether a client, gateway, or aggregation node) in an Edge computing system or like forms of appliances, computers, subsystems, circuitry, or other components.
In a more detailed example,
The Edge computing device 650 may include processing circuitry in the form of a processor 652, which may be a microprocessor, a multi-core processor, a multithreaded processor, an ultra-low voltage processor, an embedded processor, an x-PU/DPU/IPU/NPU, special purpose processing unit, specialized processing unit, or other known processing elements. The processor 652 may be a part of a system on a chip (SoC) in which the processor 652 and other components are formed into a single integrated circuit, or a single package, such as the Edison™ or Galileo™ SoC boards from Intel Corporation, Santa Clara, Calif. As an example, the processor 652 may include an Intel® Architecture Core™ based CPU processor, such as a Quark™, an Atom™, an i3, an i5, an i7, an i9, or an MCU-class processor, or another such processor available from Intel®. However, any number other processors may be used, such as available from Advanced Micro Devices, Inc. (AMD®) of Sunnyvale, Calif., a MIPS®-based design from MIPS Technologies, Inc. of Sunnyvale, Calif., an ARM®-based design licensed from ARM Holdings, Ltd. or a customer thereof, or their licensees or adopters. The processors may include units such as an A5-A13 processor from Apple® Inc., a Snapdragon™ processor from Qualcomm® Technologies, Inc., or an OMAP™ processor from Texas Instruments, Inc. The processor 652 and accompanying circuitry may be provided in a single socket form factor, multiple socket form factor, or a variety of other formats, including in limited hardware configurations or configurations that include fewer than all elements shown in
The processor 652 may communicate with a system memory 654 over an interconnect 656 (e.g., a bus). Any number of memory devices may be used to provide for a given amount of system memory. As examples, the memory 654 may be random access memory (RAM) in accordance with a Joint Electron Devices Engineering Council (JEDEC) design such as the DDR or mobile DDR standards (e.g., LPDDR, LPDDR2, LPDDR3, or LPDDR4). In particular examples, a memory component may comply with a DRAM standard promulgated by JEDEC, such as JESD79F for DDR SDRAM, JESD79-2F for DDR2 SDRAM, JESD79-3F for DDR3 SDRAM, JESD79-4A for DDR4 SDRAM, JESD209 for Low Power DDR (LPDDR), JESD209-2 for LPDDR2, JESD209-3 for LPDDR3, and JESD209-4 for LPDDR4. Such standards (and similar standards) may be referred to as DDR-based standards and communication interfaces of the storage devices that implement such standards may be referred to as DDR-based interfaces. In various implementations, the individual memory devices may be of any number of different package types such as single die package (SDP), dual die package (DDP) or quad die package (Q17P). These devices, in some examples, may be directly soldered onto a motherboard to provide a lower profile solution, while in other examples the devices are configured as one or more memory modules that in turn couple to the motherboard by a given connector. Any number of other memory implementations may be used, such as other types of memory modules, e.g., dual inline memory modules (DIMMs) of different varieties including but not limited to microDIMMs or MiniDIMMs.
To provide for persistent storage of information such as data, applications, operating systems and so forth, a storage 658 may also couple to the processor 652 via the interconnect 656. In an example, the storage 658 may be implemented via a solid-state disk drive (SSDD). Other devices that may be used for the storage 658 include flash memory cards, such as Secure Digital (SD) cards, microSD cards, eXtreme Digital (XD) picture cards, and the like, and Universal Serial Bus (USB) flash drives. In an example, the memory device may be or may include memory devices that use chalcogenide glass, multi-threshold level NAND flash memory, NOR flash memory, single or multi-level Phase Change Memory (PCM), a resistive memory, nanowire memory, ferroelectric transistor random access memory (FeTRAM), anti-ferroelectric memory, magnetoresistive random access memory (MRAM) memory that incorporates memristor technology, resistive memory including the metal oxide base, the oxygen vacancy base and the conductive bridge Random Access Memory (CB-RAM), or spin transfer torque (STT)-MRAM, a spintronic magnetic junction memory based device, a magnetic tunneling junction (MTJ) based device, a DW (Domain Wall) and SOT (Spin Orbit Transfer) based device, a thyristor based memory device, or a combination of any of the above, or other memory.
In low power implementations, the storage 658 may be on-die memory or registers associated with the processor 652. However, in some examples, the storage 658 may be implemented using a micro hard disk drive (HDD). Further, any number of new technologies may be used for the storage 658 in addition to, or instead of, the technologies described, such resistance change memories, phase change memories, holographic memories, or chemical memories, among others.
The components may communicate over the interconnect 656. The interconnect 656 may include any number of technologies, including industry standard architecture (ISA), extended ISA (EISA), peripheral component interconnect (PCI), peripheral component interconnect extended (PCIx), PCI express (PCIe), or any number of other technologies. The interconnect 656 may be a proprietary bus, for example, used in an SoC based system. Other bus systems may be included, such as an Inter-Integrated Circuit (I2C) interface, a Serial Peripheral Interface (SPI) interface, point to point interfaces, and a power bus, among others.
The interconnect 656 may couple the processor 652 to a transceiver 666, for communications with the connected Edge devices 662. The transceiver 666 may use any number of frequencies and protocols, such as 2.4 Gigahertz (GHz) transmissions under the IEEE 802.15.4 standard, using the Bluetooth® low energy (BLE) standard, as defined by the Bluetooth® Special Interest Group, or the ZigBee® standard, among others. Any number of radios, configured for a particular wireless communication protocol, may be used for the connections to the connected Edge devices 662. For example, a wireless local area network (WLAN) unit may be used to implement Wi-Fi® communications in accordance with the Institute of Electrical and Electronics Engineers (IEEE) 802.11 standard. In addition, wireless wide area communications, e.g., according to a cellular or other wireless wide area protocol, may occur via a wireless wide area network (WWAN) unit.
The wireless network transceiver 666 (or multiple transceivers) may communicate using multiple standards or radios for communications at a different range. For example, the Edge computing node 650 may communicate with close devices, e.g., within about 10 meters, using a local transceiver based on Bluetooth Low Energy (BLE), or another low power radio, to save power. More distant connected Edge devices 662, e.g., within about 50 meters, may be reached over ZigBee® or other intermediate power radios. Both communications techniques may take place over a single radio at different power levels or may take place over separate transceivers, for example, a local transceiver using BLE and a separate mesh transceiver using ZigBee®.
A wireless network transceiver 666 (e.g., a radio transceiver) may be included to communicate with devices or services in a cloud (e.g., an Edge cloud 695) via local or wide area network protocols. The wireless network transceiver 666 may be a low-power wide-area (LPWA) transceiver that follows the IEEE 802.15.4, or IEEE 802.15.4g standards, among others. The Edge computing node 650 may communicate over a wide area using LoRaWAN™ (Long Range Wide Area Network) developed by Semtech and the LoRa Alliance. The techniques described herein are not limited to these technologies but may be used with any number of other cloud transceivers that implement long range, low bandwidth communications, such as Sigfox, and other technologies. Further, other communications techniques, such as time-slotted channel hopping, described in the IEEE 802.15.4e specification may be used.
Any number of other radio communications and protocols may be used in addition to the systems mentioned for the wireless network transceiver 666, as described herein. For example, the transceiver 666 may include a cellular transceiver that uses spread spectrum (SPA/SAS) communications for implementing high-speed communications. Further, any number of other protocols may be used, such as Wi-Fi® networks for medium speed communications and provision of network communications. The transceiver 666 may include radios that are compatible with any number of 3GPP (Third Generation Partnership Project) specifications, such as Long Term Evolution (LTE) and 5th Generation (5G) communication systems, discussed in further detail at the end of the present disclosure. A network interface controller (NIC) 668 may be included to provide a wired communication to nodes of the Edge cloud 695 or to other devices, such as the connected Edge devices 662 (e.g., operating in a mesh). The wired communication may provide an Ethernet connection or may be based on other types of networks, such as Controller Area Network (CAN), Local Interconnect Network (LIN), DeviceNet, ControlNet, Data Highway+, PROFIBUS, or PROFINET, among many others. An additional NIC 668 may be included to enable connecting to a second network, for example, a first NIC 668 providing communications to the cloud over Ethernet, and a second NIC 668 providing communications to other devices over another type of network.
Given the variety of types of applicable communications from the device to another component or network, applicable communications circuitry used by the device may include or be embodied by any one or more of components 664, 666, 668, or 670. Accordingly, in various examples, applicable means for communicating (e.g., receiving, transmitting, etc.) may be embodied by such communications circuitry.
The Edge computing node 650 may include or be coupled to acceleration circuitry 664, which may be embodied by one or more artificial intelligence (AI) accelerators, a neural compute stick, neuromorphic hardware, an FPGA, an arrangement of GPUs, an arrangement of x-PUs/DPUs/IPU/NPUs, one or more SoCs, one or more CPUs, one or more digital signal processors, dedicated ASICs, or other forms of specialized processors or circuitry designed to accomplish one or more specialized tasks. These tasks may include AI processing (including machine learning, training, inferencing, and classification operations), visual data processing, network data processing, object detection, rule analysis, or the like. These tasks also may include the specific Edge computing tasks for service management and service operations discussed elsewhere in this document.
The interconnect 656 may couple the processor 652 to a sensor hub or external interface 670 that is used to connect additional devices or subsystems. The devices may include sensors 672, such as accelerometers, level sensors, flow sensors, optical light sensors, camera sensors, temperature sensors, global navigation system (e.g., GPS) sensors, pressure sensors, barometric pressure sensors, and the like. The hub or interface 670 further may be used to connect the Edge computing node 650 to actuators 674, such as power switches, valve actuators, an audible sound generator, a visual warning device, and the like.
In some optional examples, various input/output (I/O) devices may be present within or connected to, the Edge computing node 650. For example, a display or other output device 684 may be included to show information, such as sensor readings or actuator position. An input device 686, such as a touch screen or keypad may be included to accept input. An output device 684 may include any number of forms of audio or visual display, including simple visual outputs such as binary status indicators (e.g., light-emitting diodes (LEDs)) and multi-character visual outputs, or more complex outputs such as display screens (e.g., liquid crystal display (LCD) screens), with the output of characters, graphics, multimedia objects, and the like being generated or produced from the operation of the Edge computing node 650. A display or console hardware, in the context of the present system, may be used to provide output and receive input of an Edge computing system; to manage components or services of an Edge computing system; identify a state of an Edge computing component or service; or to conduct any other number of management or administration functions or service use cases.
A battery 676 may power the Edge computing node 650, although, in examples in which the Edge computing node 650 is mounted in a fixed location, it may have a power supply coupled to an electrical grid, or the battery may be used as a backup or for temporary capabilities. The battery 676 may be a lithium ion battery, or a metal-air battery, such as a zinc-air battery, an aluminum-air battery, a lithium-air battery, and the like.
A battery monitor/charger 678 may be included in the Edge computing node 650 to track the state of charge (SoCh) of the battery 676, if included. The battery monitor/charger 678 may be used to monitor other parameters of the battery 676 to provide failure predictions, such as the state of health (SoH) and the state of function (SoF) of the battery 676. The battery monitor/charger 678 may include a battery monitoring integrated circuit, such as an LTC4020 or an LTC2990 from Linear Technologies, an ADT7488A from ON Semiconductor of Phoenix Ariz., or an IC from the UCD90xxx family from Texas Instruments of Dallas, Tex. The battery monitor/charger 678 may communicate the information on the battery 676 to the processor 652 over the interconnect 656. The battery monitor/charger 678 may also include an analog-to-digital (ADC) converter that enables the processor 652 to directly monitor the voltage of the battery 676 or the current flow from the battery 676. The battery parameters may be used to determine actions that the Edge computing node 650 may perform, such as transmission frequency, mesh network operation, sensing frequency, and the like.
A power block 680, or other power supply coupled to a grid, may be coupled with the battery monitor/charger 678 to charge the battery 676. In some examples, the power block 680 may be replaced with a wireless power receiver to obtain the power wirelessly, for example, through a loop antenna in the Edge computing node 650. A wireless battery charging circuit, such as an LTC4020 chip from Linear Technologies of Milpitas, Calif., among others, may be included in the battery monitor/charger 678. The specific charging circuits may be selected based on the size of the battery 676, and thus, the current required. The charging may be performed using the Airfuel standard promulgated by the Airfuel Alliance, the Qi wireless charging standard promulgated by the Wireless Power Consortium, or the Rezence charging standard, promulgated by the Alliance for Wireless Power, among others.
The storage 658 may include instructions 682 in the form of software, firmware, or hardware commands to implement the techniques described herein. Although such instructions 682 are shown as code blocks included in the memory 654 and the storage 658, it may be understood that any of the code blocks may be replaced with hardwired circuits, for example, built into an application specific integrated circuit (ASIC).
In an example, the instructions 682 provided via the memory 654, the storage 658, or the processor 652 may be embodied as a non-transitory, machine-readable medium 660 including code to direct the processor 652 to perform electronic operations in the Edge computing node 650. The processor 652 may access the non-transitory, machine-readable medium 660 over the interconnect 656. For instance, the non-transitory, machine-readable medium 660 may be embodied by devices described for the storage 658 or may include specific storage units such as storage devices and/or storage disks that include optical disks (e.g., digital versatile disk (DVD), compact disk (CD), CD-ROM, Blu-ray disk), flash drives, floppy disks, hard drives (e.g., SSDs), or any number of other hardware devices in which information is stored for any duration (e.g., for extended time periods, permanently, for brief instances, for temporarily buffering, and/or caching). The non-transitory, machine-readable medium 660 may include instructions to direct the processor 652 to perform a specific sequence or flow of actions, for example, as described with respect to the flowchart(s) and block diagram(s) of operations and functionality depicted above. As used herein, the terms “machine-readable medium” and “computer-readable medium” are interchangeable. As used herein, the term “non-transitory computer-readable medium” is expressly defined to include any type of computer readable storage device and/or storage disk and to exclude propagating signals and to exclude transmission media.
Also in a specific example, the instructions 682 on the processor 652 (separately, or in combination with the instructions 682 of the machine readable medium 660) may configure execution or operation of a trusted execution environment (TEE) 690. In an example, the TEE 690 operates as a protected area accessible to the processor 652 for secure execution of instructions and secure access to data. Various implementations of the TEE 690, and an accompanying secure area in the processor 652 or the memory 654 may be provided, for instance, through use of Intel® Software Guard Extensions (SGX) or ARM® TrustZone® hardware security extensions, Intel® Management Engine (ME), or Intel® Converged Security Manageability Engine (CSME). Other aspects of security hardening, hardware roots-of-trust, and trusted or protected operations may be implemented in the device 650 through the TEE 690 and the processor 652.
While the illustrated examples of
In some examples, computers operating in a distributed computing and/or distributed networking environment (e.g., an Edge network) are structured to accommodate particular objective functionality in a manner that reduces computational waste. For instance, because a computer includes a subset of the components disclosed in
In the illustrated examples of
The instructions 682 may further be transmitted or received over a communications network using a transmission medium via the wireless network transceiver 466 utilizing any one of a number of wireless local area network (WLAN) transfer protocols (e.g., frame relay, internet protocol (IP), transmission control protocol (TCP), user datagram protocol (UDP), hypertext transfer protocol (HTTP), etc.). Example communication networks may include a local area network (LAN), a wide area network (WAN), a packet data network (e.g., the Internet), mobile telephone networks (e.g., cellular networks), Plain Old Telephone (POTS) networks, and wireless data networks. Communications over the networks may include one or more different protocols, such as Institute of Electrical and Electronics Engineers (IEEE) 802.11 family of standards known as Wi-Fi, IEEE 802.16 family of standards, IEEE 802.15.4 family of standards, a Long Term Evolution (LTE) family of standards, a Universal Mobile Telecommunications System (UMTS) family of standards, peer-to-peer (P2P) networks, a next generation (NG)/5th generation (5G) standards among others.
Note that the term “circuitry” as used herein refers to, is part of, or includes hardware components such as an electronic circuit, a logic circuit, a processor (shared, dedicated, or group) and/or memory (shared, dedicated, or group), an Application Specific Integrated Circuit (ASIC), a field-programmable device (FPD) (e.g., a field-programmable gate array (FPGA), a programmable logic device (PLD), a complex PLD (CPLD), a high-capacity PLD (HCPLD), a structured ASIC, or a programmable SoC), digital signal processors (DSPs), etc., that are configured to provide the described functionality. In some embodiments, the circuitry may execute one or more software or firmware programs to provide at least some of the described functionality. The term “circuitry” may also refer to a combination of one or more hardware elements (or a combination of circuits used in an electrical or electronic system) with the program code used to carry out the functionality of that program code. In these embodiments, the combination of hardware elements and program code may be referred to as a particular type of circuitry.
The term “processor circuitry” or “processor” as used herein thus refers to, is part of, or includes circuitry capable of sequentially and automatically carrying out a sequence of arithmetic or logical operations, or recording, storing, and/or transferring digital data. The term “processor circuitry” or “processor” may refer to one or more application processors, one or more baseband processors, a physical central processing unit (CPU), a single- or multi-core processor, and/or any other device capable of executing or otherwise operating computer-executable instructions, such as program code, software modules, and/or functional processes.
Any of the radio links described herein may operate according to any one or more of the following radio communication technologies and/or standards including but not limited to: a Global System for Mobile Communications (GSM) radio communication technology, a General Packet Radio Service (GPRS) radio communication technology, an Enhanced Data Rates for GSM Evolution (EDGE) radio communication technology, and/or a Third Generation Partnership Project (3GPP) radio communication technology, for example Universal Mobile Telecommunications System (UMTS), Freedom of Multimedia Access (FOMA), 3GPP Long Term Evolution (LTE), 3GPP Long Term Evolution Advanced (LTE Advanced), Code division multiple access 2000 (CDMA2000), Cellular Digital Packet Data (CDPD), Mobitex, Third Generation (3G), Circuit Switched Data (CSD), High-Speed Circuit-Switched Data (HSCSD), Universal Mobile Telecommunications System (Third Generation) (UMTS (3G)), Wideband Code Division Multiple Access (Universal Mobile Telecommunications System) (W-CDMA (UMTS)), High Speed Packet Access (HSPA), High-Speed Downlink Packet Access (HSDPA), High-Speed Uplink Packet Access (HSUPA), High Speed Packet Access Plus (HSPA+), Universal Mobile Telecommunications System-Time-Division Duplex (UMTS-TDD), Time Division-Code Division Multiple Access (TD-CDMA), Time Division-Synchronous Code Division Multiple Access (TD-CDMA), 3rd Generation Partnership Project Release 8 (Pre-4th Generation) (3GPP Rel. 8 (Pre-4G)), 3GPP Rel. 9 (3rd Generation Partnership Project Release 9), 3GPP Rel. 10 (3rd Generation Partnership Project Release 10), 3GPP Rel. 11 (3rd Generation Partnership Project Release 11), 3GPP Rel. 12 (3rd Generation Partnership Project Release 12), 3GPP Rel. 13 (3rd Generation Partnership Project Release 13), 3GPP Rel. 14 (3rd Generation Partnership Project Release 14), 3GPP Rel. 15 (3rd Generation Partnership Project Release 15), 3GPP Rel. 16 (3rd Generation Partnership Project Release 16), 3GPP Rel. 17 (3rd Generation Partnership Project Release 17) and subsequent Releases (such as Rel. 18, Rel. 19, etc.), 3GPP 5G, 5G, 5G New Radio (5G NR), 3GPP 5G New Radio, 3GPP LTE Extra, LTE-Advanced Pro, LTE Licensed-Assisted Access (LAA), MuLTEfire, UMTS Terrestrial Radio Access (UTRA), Evolved UMTS Terrestrial Radio Access (E-UTRA), Long Term Evolution Advanced (4th Generation) (LTE Advanced (4G)), cdmaOne (2G), Code division multiple access 2000 (Third generation) (CDMA2000 (3G)), Evolution-Data Optimized or Evolution-Data Only (EV-DO), Advanced Mobile Phone System (1st Generation) (AMPS (1G)), Total Access Communication System/Extended Total Access Communication System (TACS/ETACS), Digital AMPS (2nd Generation) (D-AMPS (2G)), Push-to-talk (PTT), Mobile Telephone System (MTS), Improved Mobile Telephone System (IMTS), Advanced Mobile Telephone System (AMTS), OLT (Norwegian for Offentlig Landmobil Telefoni, Public Land Mobile Telephony), MTD (Swedish abbreviation for Mobiltelefonisystem D, or Mobile telephony system D), Public Automated Land Mobile (Autotel/PALM), ARP (Finnish for Autoradiopuhelin, “car radio phone”), NMT (Nordic Mobile Telephony), High capacity version of NTT (Nippon Telegraph and Telephone) (Hicap), Cellular Digital Packet Data (CDPD), Mobitex, DataTAC, Integrated Digital Enhanced Network (iDEN), Personal Digital Cellular (PDC), Circuit Switched Data (CSD), Personal Handy-phone System (PHS), Wideband Integrated Digital Enhanced Network (WiDEN), iBurst, Unlicensed Mobile Access (UMA), also referred to as also referred to as 3GPP Generic Access Network, or GAN standard), Zigbee, Bluetooth®, Wireless Gigabit Alliance (WiGig) standard, mmWave standards in general (wireless systems operating at 10-300 GHz and above such as WiGig, IEEE 802.11ad, IEEE 802.11ay, etc.), technologies operating above 300 GHz and THz bands, (3GPP/LTE based or IEEE 802.11p or IEEE 802.11bd and other) Vehicle-to-Vehicle (V2V) and Vehicle-to-X (V2X) and Vehicle-to-Infrastructure (V2I) and Infrastructure-to-Vehicle (I2V) communication technologies, 3GPP cellular V2X, DSRC (Dedicated Short Range Communications) communication systems such as Intelligent-Transport-Systems and others (typically operating in 5850 MHz to 5925 MHz or above (typically up to 5935 MHz following change proposals in CEPT Report 71)), the European ITS-G5 system (i.e. the European flavor of IEEE 802.11p based DSRC, including ITS-G5A (i.e., Operation of ITS-G5 in European ITS frequency bands dedicated to ITS for safety related applications in the frequency range 5,875 GHz to 5,905 GHz), ITS-G5B (i.e., Operation in European ITS frequency bands dedicated to ITS non-safety applications in the frequency range 5,855 GHz to 5,875 GHz), ITS-G5C (i.e., Operation of ITS applications in the frequency range 5,470 GHz to 5,725 GHz)), DSRC in Japan in the 700 MHz band (including 715 MHz to 725 MHz), IEEE 802.11bd based systems, etc.
Aspects described herein can be used in the context of any spectrum management scheme including dedicated licensed spectrum, unlicensed spectrum, license exempt spectrum, (licensed) shared spectrum (such as LSA=Licensed Shared Access in 2.3-2.4 GHz, 3.4-3.6 GHz, 3.6-3.8 GHz and further frequencies and SAS=Spectrum Access System/CBRS=Citizen Broadband Radio System in 3.55-3.7 GHz and further frequencies). Applicable spectrum bands include IMT (International Mobile Telecommunications) spectrum as well as other types of spectrum/bands, such as bands with national allocation (including 450-470 MHz, 902-928 MHz (note: allocated for example in the US (FCC Part 15)), 863-868.6 MHz (note: allocated for example in European Union (ETSI EN 300 220)), 915.9-929.7 MHz (note: allocated for example in Japan), 917-923.5 MHz (note: allocated for example in South Korea), 755-779 MHz and 779-787 MHz (note: allocated for example in China), 790-960 MHz, 1710-2025 MHz, 2110-2200 MHz, 2300-2400 MHz, 2.4-2.4835 GHz (note: it is an ISM band with global availability and it is used by Wi-Fi technology family (11b/g/n/ax) and also by Bluetooth), 2500-2690 MHz, 698-790 MHz, 610-790 MHz, 3400-3600 MHz, 3400-3800 MHz, 3800-4200 MHz, 3.55-3.7 GHz (note: allocated for example in the US for Citizen Broadband Radio Service), 5.15-5.25 GHz and 5.25-5.35 GHz and 5.47-5.725 GHz and 5.725-5.85 GHz bands (note: allocated for example in the US (FCC part 15), consists four U-NII bands in total 500 MHz spectrum), 5.725-5.875 GHz (note: allocated for example in EU (ETSI EN 301 893)), 5.47-5.65 GHz (note: allocated for example in South Korea, 5925-7085 MHz and 5925-6425 MHz band (note: under consideration in US and EU, respectively. Next generation Wi-Fi system is expected to include the 6 GHz spectrum as operating band, but it is noted that, as of December 2017, Wi-Fi system is not yet allowed in this band. Regulation is expected to be finished in 2019-2020 time frame), IMT-advanced spectrum, IMT-2020 spectrum (expected to include 3600-3800 MHz, 3800-4200 MHz, 3.5 GHz bands, 700 MHz bands, bands within the 24.25-86 GHz range, etc.), spectrum made available under FCC's “Spectrum Frontier” 5G initiative (including 27.5-28.35 GHz, 29.1-29.25 GHz, 31-31.3 GHz, 37-38.6 GHz, 38.6-40 GHz, 42-42.5 GHz, 57-64 GHz, 71-76 GHz, 81-86 GHz and 92-94 GHz, etc.), the ITS (Intelligent Transport Systems) band of 5.9 GHz (typically 5.85-5.925 GHz) and 63-64 GHz, bands currently allocated to WiGig such as WiGig Band 1 (57.24-59.40 GHz), WiGig Band 2 (59.40-61.56 GHz) and WiGig Band 3 (61.56-63.72 GHz) and WiGig Band 4 (63.72-65.88 GHz), 57-64/66 GHz (note: this band has near-global designation for Multi-Gigabit Wireless Systems (MGWS)/WiGig. In US (FCC part 15) allocates total 14 GHz spectrum, while EU (ETSI EN 302 567 and ETSI EN 301 217-2 for fixed P2P) allocates total 9 GHz spectrum), the 70.2 GHz-71 GHz band, any band between 65.88 GHz and 71 GHz, bands currently allocated to automotive radar applications such as 76-81 GHz, and future bands including 94-300 GHz and above. Furthermore, the scheme can be used on a secondary basis on bands such as the TV White Space bands (typically below 790 MHz) where in particular the 400 MHz and 700 MHz bands are promising candidates. Besides cellular applications, specific applications for vertical markets may be addressed such as PMSE (Program Making and Special Events), medical, health, surgery, automotive, low-latency, drones, etc. applications.
In the illustrated example of
In the illustrated example of
With this type of serverless data center 800, a new way to perform orchestration may be achieved, which moves from current practices to an objective-driven approach where the customer expresses only the intent and the orchestration stack itself sets up the platform to achieve that intent.
Some existing Software as a Service (SaaS) models typically promote to their consumers a service-specific set of Service Level Objectives (SLO). For instance, a data analytics platform may promote a number of jobs per hour that can be computed. This is great from a user perspective, but still requires the SaaS provider to preemptively map SLOs to resources needed before sending requests to the resource orchestrator. The service providers are not able to automatically select resource types based on SLOs.
The systems and methods described herein illustrate how intents (objectives) may be mapped to platform settings automatically, dynamically, and directly. Intents are received as higher-level objectives and mapped throughout the orchestration stack to lower-level settings.
For example, consider an application requirement that there is certain percentage that something completes in a time window, referred to herein as “P50 Latency.” Other terms may also be used, which may be similar, such as “P50 Target” or “P50 Completions” to indicate that tasks, projects, requests, or the like, need to complete on time at least 50% of the time, or that there is at least a 50% probability that the task will complete on time. This intent is mapped to lower-level settings, such as to indicate a thread-level priority and from there to a CPU cache way/profile assignment/resource allocation. On the input/output (I/O) side, the intent may be mapped to network settings to ensure sufficient communication resources are available to transmit and receive the request and response.
To dynamically adapt to changing conditions, the systems and methods use control loops at various levels in the architecture. The control loops may employ credit systems, utility/cost functions, planning and solvers, etc. The control loops may execute at different speeds depending on how quickly the monitored system needs to adapt. Such control loops are especially important and useful when considering the dynamicity of compute systems. For instance, network traffic may vary daily. If customers have to change the resource requirements themselves, then there is additional burden on them to use the service provider. Additionally, such changes may be difficult because the information needed to make such decisions may not be exposed to the customer. Instead, using the systems and methods described here, the customer need only to specify the intent and the system automatically adjusts the resources. This may lead to cost savings for both the customer and the service provider.
Dynamicity in the system also applies to the plans that are generated. These plans have a temporal element. For example, a plan may temporarily (e.g., on the order of minutes) target allocating an abundance of resources to a workload to catch up with an SLA under the supervision of a fast control loop and then over a longer period (e.g., monthly) look to achieve an acceptable and affordable resource allocation pattern with the supervision of a slower control loop. The plan can be either automatically triggered or be partially manually triggered, for instance by first sending the plan as a recommendation to a human operator to enable human-guidance.
Additionally, a plan's temporal aspect can also be useful from a continuous improvement point of view. A plan that satisfies all SLOs in the system-of-systems might be replaced by another plan which also satisfies all SLOs but triggers a different setup, configuration, or set of policies. For instance, to prepare for maintenance, an SLO may use resource more efficiently, make room for incoming workloads/services, or the like.
With a decentralized service-oriented platform, the complexity of orchestration increases even more. The systems and methods described here address orchestration across multiple sites that may be owned by different stakeholders. The individual parts may implement negotiation and coordination to achieve an overarching intent, which is paramount when there is no centralized orchestration.
In an implementation, a service provider may deploy a unit of computing (e.g., a pod) with application context meta-data so that the node can, in a fast control loop, make more appropriate application decisions and possibly choose not to deprioritize a low priority workload because of its end-to-end impacting nature. This requires an E2E view, however, and it include notions of co-scheduling and co-prioritization of nodes that must provide for security functions over data (including data provenance, compliance, and recording) with other computational nodes that perform operations in higher layers of the stack. In Kubernetes, a Pod is the smallest deployable unit of computing that you can create and manage A Pod is a group of one or more containers, with shared storage and network resources, and a specification for how to run the containers. A pod's contents are always colocated and co-scheduled, and run in a shared context.
With the shift to intent-driven systems, the system focuses on shifting from the deployment function to whether resources could potentially be made available, and how to reconfigure the system to make them available. Resources may be of different types while being in the same category: e.g. high-performance processors (e.g., Intel® Xeon CPUs) versus several reduced-computation processors (e.g., Xeon-Ds), or a single large XPU versus several smaller XPUs. Furthermore, a resource itself may contain different types of sub-resources/components, for instance efficient and performant cores, etc.
In some implementations, the system addresses the challenge of administratively governing complex, distributed infrastructure by putting Service Level Objective oriented controls in the hands of cluster and infrastructure administrators and removing the burden of having to specify possibly erroneous low level configuration details. An Admin Policy Controller and an Admin Policy Translation module may be implemented in the orchestration architecture and a workflow to administer multiple tiers of intent driven admin policies may be used in the workload deployment workflow. A closed loop control of administrative policies, including how to achieve them with multiple iterations on imperative configurations, may be used. It can influence workload placement and re-placement in the face of changing policies or policy non-compliance.
The system may implement a new model for driving Kubernetes' life-cycle management (LCM) to reflect the needs of the application rather than the assumptions of the Kubernetes administrator. The application is deployed temporarily (to support fast start) and subsequently a new instance of Kubernetes is deployed with the cluster and node level policies that better reflects the needs of the application. The workload is subsequently moved to the new cluster.
Current methods of sharing joint resources like Cache Allocation Technology (CAT) and Memory Bandwidth Allocation (MBA), and power rely on the application's KPIs being monitored by a specific dynamic controller (e.g., Resource Management Daemon (RMD), Dynamic Resource Controller (DRC), Appqos, etc.) or using Telemetry Aware Scheduling to deal with exhaustion. Having a “resource request broker” integrated into orchestration can have a message-based system, independent of the application KPI (which is monitored by the application itself). The broker can have a local agent on the platform, since time is a factor in the share, and can make local decisions based on a cluster “share” policy. Individual pods can sign up for static contracts, “guaranteed resource contracts,” equal share contracts, or dynamic (bidding) contracts. Competing requests are arbitrated by the local broker agent which can operate a bid/offer scheme, for instance, “I'm willing to give up resource X for resource Y for period N.”
Security risks arise when components are distributed across several resources. In intent-based orchestration, and similar to QoS classes, the system may allow users to define how “sensitive” they are, or how much risk they are willing to accepts, to certain aspects of security or trust in the environment. The system may include a set of additional components to the orchestration control plane to monitor, analyze, and access risk as well as a component able to translate the assessment into a set of policies, resources allocations, etc.
SLOs, either full-stack or E2E (e.g., system to system) may be varied over time, be expressed with a range (e.g., a minimum and maximum allowable value), allow for deviance or variance outside of ranges for a certain period of time, be expressed as preferred or using other prioritization schemes, or the like. An example may be that the task must be P99 compliant 99% of the time for 60 minutes while not exceeding 10 ms and running at a preferred compliance level of 5 ms.
In some implementations, a set of sidecars associated with their workload are used on a node. The sidecars coordinate to make sure that the node, platform, facilities, etc. are setup correctly so the objectives of their workload can be met. The sidecars can monitor a particular setup and enforce the correct policies, allocating the correct resources, tuning settings so that the intents of the pods can be met, etc.
In some installations, there exists a challenge of placing charging/billing constraints on resource allocation decided upon by the orchestration system when presented with Service Level Objective based input parameters for a workload. Two new components enable Intent Driven Charging and Intent Driven Billing: 1) Charging Guardrails Function and 2) SLO Planning Function. The concept of a Charging Guardrails Function is introduced into the deployment workflow that is a logically central point in the orchestration architecture. It acts as the entity responsible for controlling and guiding the SLO oriented resource planning component to make sure the allocated resources are affordable for the user. The SLO Planning Function needs to consider the cost basis for allocated resources and not just their suitability for likely adherence to the workload SLA.
The planner 1012 may be implemented using a hardwired circuit, a programmable hardware device (e.g., an ASIC), or as instructions executing on a hardware platform (e.g., on a general CPU). The planner 1012 may be configured, designed, programmed, or otherwise adapted to map intents or objectives to SLOs. The planner 1012 may also break down SLOs into actionable plans. The planner 1012 may be used to automatically translate or map SLO requirements to appropriate pod specifications, which may include resource requirements, platform features, or policies across compute, network, storage, and facilities (e.g., power) domains. Mapping of objectives to policies and lower-level objective settings may implement heuristic, machine learning (ML), or artificial intelligence (AI) mechanisms, workload characterizations (e.g., derived from online data or through offline experimentation, or through sandboxing), or policies from resource owners to guide how to use the owner's system.
The planner 1012 may further be used to coordinate system-to-system (e.g., E2E) with another planner 1022 and map multiple points-of-presence (PoPs) belonging to potentially different stakeholders. Various types of coordination and negotiation schemes may be used. For instance, a multiple attribute utility theory (MAUT) model may be used to negotiate resource allocation.
The planner 1012 oversees translating intents into actional policies, system settings, resource requirements, and the like. It does this based on insights stored in the knowledge database 1020. Once a plan is available, it is implemented and enforced by the schedulers 1010.
A plan may be defined using multiple SLOs at various levels in the orchestration control plane 1000. For instance, SLOs on higher levels of the orchestration control plane 1000 may be used to regulate and control FPS and SLOs on lower levels of the orchestration control plane 1000 may be used to regulate and control IPC. Vendors, users, and service providers may define SLOs using a standardized format. SLOs may also include guardrails, which provide some variation from the target or limit values. For instance, a guardrail may allow for 10% violation of P95 for up to 10 minutes. Guardrails may be conditional as well. For instance, a guardrail may allow for 10% violation of P95 for up to 10 minutes if the system will guarantee P99 compliance thereafter.
With increasing heterogenous compute setups and XPU environments, the lower-level orchestration and resource managers may be configured to support co-scheduling and co-prioritization of workloads at a very granular level.
When assigning tasks to resources, the placement problem is difficult to solve. Unlike cluster-based scheduling, this system allows local resources to be reprioritized based on the closeness of the operation to required completion deadline. Load balance and auto-scale actions initiated locally at graph bottlenecks without having to be driven from a centralized scheduler. These actions are also time limited, as compared to durable instructions from a centralized or hierarchical scheduler.
Some resources, such as processors, may be orchestrated in a manner to optimize power usage. Processor power levels (e.g., PL1, PL2, PL3, etc.) are used to define the threshold power draw. The system platform continually adjusts these PL values and other power values (capping, P state, uncore frequency) in a fine-grained manner But because this is a complex mapping given the highly time-varying, situational nature of transfer function linking performance to power to SLAs, the adjustments are driven through pre-trained models. The pre-trained models take as inputs various timeseries derived trend signals over utilizations and power, and they may in general be models that are trained on a coarse-grained types of workload mixes (e.g., “AI”, “gRPC heavy”, “media”, etc.).
When the customer, expresses storage access intents when registering functions, these registration intents provide sufficient context to the infrastructure layer to enable efficient data movement from storage layer to the function. This is done by either scheduling the function right next to the data, using compute accelerators on servers holding the data or utilizing efficient data transfer mechanisms within or across servers. The key point is that this is achieved without the function needing specific knowledge of the server location or having to program specific compute resources or transfer mechanisms. Scheduling/placement is a function at the start of a process. As access patterns may not be obvious at deployment time, this system also implements a runtime assessment of data locality and transfer rates and using that insight to trigger auto rescheduling events, for instance, histograms that show which function/microservice is good at processing what kind of data.
The CIM 1016 scouts for options to make the current operations more efficient, while analyzing trade-offs or changes in policy, and is in charge of predicting how the system will behave in near future. This prediction is key to enable proactive planning Either reactive or proactive planning may be used with various time scales (to satisfy the various fast/slow) loops.
The monitor 1014 takes inputs from the observability stack 1018 and uses this information to feed or trigger the planner 1012 and CIM 1016. Furthermore, observability stack 1018 is in charge of assuring the insights, heuristics, etc., in the knowledge database 1020 are accurate using online and offline learning processes. The offline learning may be achieved through experiment-based workload characterization. The observability stack 1018 may collect training data to make the architecture self-evolving through analytics that use the collected data. The training data may be derived from real-world data that was captured by monitors. Alternatively, the training data may be prepared offline and provided to the observability stack 1018. Use of a CIM 1016 provides the advantage of having continuous improvement capabilities and enabling self-adaptation, self-healing, and optimization in the system.
To provide continuous improvement, multiple control loops may be implemented.
The SLO translator 1102 can use a database of rules and dictionaries (e.g., knowledge database 1020) to map the intent-based SLO to three aspects: 1) monitoring parameters, 2) configuration parameters, and 3) time-domain parameters.
The monitoring parameters are used by the service monitor 1104 when monitoring services, tasks, or jobs. Monitoring parameters may include a series of flexible guardrails and required telemetry for the service monitor 1104 to actively monitor the operations.
Guardrails may provide time-bounded and range-bounded flexibility, so that they become context sensitive and situation adaptive. The guardrails that apply on an orchestrator's function may thus shift very slowly but seamlessly to permit maximizing a desirable end-to-end objective by making a series of localized leeways. The idea is to achieve three important flexibilities: a) take the hardest constraint—such as P99.9 latency—and allow it to move from a fixed threshold to a threshold range, say +10 percent for a short duration in time, to be compensated by an even better P99.9 latency for an extended period of time later, b) take a possibly costly adjustment in resource allocation at one time, and move it out by a bounded amount of time, and in the interim, soften the guardrails so that satisfiable solutions can continue to run, and c) permit a richer system response to emergency needs, demands, etc., to deal with transient failures in other parts of a data center which require resources to be shared more flexibly until repair and subsequent normal operations can pick up.
Current methods for guardrails involve using thresholding and integration over a pre-defined period to evaluate a guardrail crossing and subsequent remediation. Allowing guardrails to become “stateful” or “context” based allows hard constraints to be softened during specific conditions. For example, CPU utilization maybe allowed to peak on first deployment of workload as the workload caches data but after that time, in normal operation, excessive CPU utilization can trigger a guardrail crossing. The system adds knowledge of context including life cycle staging, in service upgrades, HA failovers, service mesh restarts, etc. as contextual information for implementing guardrails.
The configuration parameters are used by orchestrators and resource managers to configure compute, network, memory, and other resources. The configuration parameters may be expressed as a series of orchestration objectives, which are fed to the highest-level orchestration system and turned into a series of sub-objectives by each lower layer of the orchestration stack. At the bottom of the orchestration stack are physical hardware components that may be tuned with fine-grained control. As such, a series of resource controllers may enforce policies on the compute platform. The compute platform may be a CPU, GPU, FPGA, accelerator, IPU, or other type of processing device.
The time-domain parameters are used to configure the control loops to set monitoring cycles and how often changes are made to configuration parameters. The SLO translator 1102 generates time domains for SLO monitoring, which range from non-real-time monitoring to real-time monitoring. The time domains specify strict monitoring and orchestration feedback response times for the corresponding SLO. The time domains are shown in
The service monitor 1104 has layers that monitor E2E telemetry 1106, non-real-time telemetry 1108, near real-time telemetry or real-time telemetry 1110. Each layer has a corresponding control loop, which may have differing time-domain parameters.
The SLO translator 1102 converts the intents into “Service level monitoring” for E2E, slow resource monitoring, and fast resource monitoring. Based on rules, policies, and learned insights, the SLO translator 1102 maps the intent to one or more service monitors, which may be instantiated based on the type of intent, the reaction speed required, or other requirements. The SLO translator 1102 configures the service monitors to provide notifications to entities in the orchestration stack when the “physical SLAs” or “physical SLOs” are out of bounds. Service monitors can include classic service assurance solutions to monitor E2E SLAs using passive or active probes, software defined networking (SDN) controllers, or SDN analytics systems. Faster service monitors can be colocated on the platforms, as needed, to achieve the response time required by the mapped SLO.
The system may employ methods to inject smart observability, for instance, on the deployment of components, and how to auto-inject the monitoring at the correct location. After deployment of the components, the system may form a part of the control loop that achieves automatic SLW remediation. When there is a case for remediation, the system ensures that sufficient controls are in place to provide a corrective action that is effective and does not impact other services inadvertently.
As illustrated in
In contrast to standard imperative orchestration, intent-based orchestration can be enabled by allowing the user to describe, as part of a component's requirements, an intent. This is a declarative mechanism where the user is declaring a desired outcome. So, instead of expressing specific rules or parameters (such as in imperative mechanisms), the user is able to express a desired result. An example might be to achieve some given level of availability or maximum processing latency, for which multiple instances may be deployed on nodes with specific characteristics to achieve that intent. An intent is a declarative expression.
Rather than deeply integrating intent-based orchestration into a standard orchestrator's scheduler, an intent-based scheduling algorithm may be deployed in a sub-orchestrator. In this case, when the top-level standard orchestrator receives the application description, it may see an intent specified for one or more components of the application. It may choose to request an intent-based sub-orchestrator to do the scheduling of these components. Each intent-based orchestrator or sub-orchestrator may specialize in meeting specific kinds of intents. The intent-based sub-orchestrator may further break down the problem into other sub-orchestrators.
For example, consider a video analytics pipeline consisting of an ingest step, a processing step, and an actuation step. The overall application may be described as an ingest component for each camera, a processing step with an intent of no more than 100 ms of latency, and an actuation component for each actuator. The top-level orchestrator can handle the ingestion and actuation component deployments. The processing could be passed to an intent-based orchestrator which might determine how many processing components are needed to load balance and achieve the desired latency. The intent-based orchestrator may even subdivide the task to additional sub-orchestrators, so that multiple clusters of nodes are used to achieve the intent (or perhaps to enable an additional intent of high availability at a cluster level).
There are several advantages of this approach. First, it is not necessary to merge the complex decision making of existing scheduling algorithms in existing orchestrators with the equally complex decision making of intent based orchestration. Each can be used for the key parts of the problem. In addition, distributed decision making allows the decision to be pushed close to the processing. This may allow for faster reactivity, which will help enable things like fast control loops in industrial use cases.
In various embodiments, a top-level orchestrator is configured, programmed, or otherwise adapted to receive intents from a customer, identify when an intent-based sub-orchestration is required, and define interactions between the top-level orchestrator and sub-orchestrators. The top-level orchestrator may be a standard orchestrator and the sub-orchestrators may be an intent-based orchestrator, such as one described in
To achieve this organization of orchestrators, intents should be expressed in a way that is compatible with standard orchestrators and these standard orchestrators should be able to identify when an intent-based sub-orchestration is required. A protocol may be used between the requesting orchestrator and sub-orchestrators that are used to satisfy the requesting orchestrator.
Further, when an application is split into components, which are separately orchestrated, they may have ordering dependencies that affect the overall orchestration. Where such ordering dependencies are present, they may be described in abstract terms. For example, in a use case of a producer-consumer flow, the producing component may be specified as desirably staying X units of data, events, frames, etc. ahead of the consuming component. Accordingly, the sub-orchestrators for each component may have conditional responsibilities for allocation of resources (consumer needing fewer resources at time T0 than at time T0+delta, while the producer needing more resources at time T0−delta than at time T0). These “resource flows” become tightly coordinated among the sub-orchestrators so that the producer and consumer in our example collectively get an X-fraction of the CPU, cache, memory bandwidth, etc., but where the resources flow seamlessly according to the choreographed sharing between them. Similarly, priority inversions need to be prevented; thus, while a producer may have a lower priority because, say the consumer is trying hard to catch up and reduce the distance by which the producer is ahead of it, if the consumer continues to cut the distance so fast that the producer now has to struggle to stay ahead, then it makes sense for the priorities to flow quickly between them as a function of the distance, instead of requiring intrusive tuning of the scheduling software.
To achieve the requested intent/objectives, the demand for acceleration technology (FPGAs, GPUs, etc.) as well as hardware features supporting the overall setup (e.g., DRC) may be implemented. Often, accelerators and XPUs run operations that a CPU code developer does not control directly though instructions. Thus, sensitive operations performed with a certain hardware privilege in an accelerator or an XPU, or in some manner opaque to ordinary software in higher levels of the stack, can simplify the security boundary constraints that would otherwise require co-scheduling of security filtering operations. Furthermore, lower-level features need to inform the various control loops of what is happening in fast control loops. This type of reporting requires extensions/usage of observability frameworks for tracing, logging, and monitoring.
At 1204, the SLO is mapped to a plurality of policies. The mapping may be based on a static map. Alternatively, the mapping may be performed using heuristics or other intelligent mechanisms.
At 1206, the plurality of policies are distributed to a plurality of sub-orchestrators, where each of the plurality of sub-orchestrators manage execution of a portion of the task. Policies may be grouped or separated by type, by resource, or by other factors. The sub-orchestrators may be responsible for a group of resources, a type of resource, a particular node or set of nodes, or the like.
At 1208, the execution of the task is monitored. Task monitoring may be performed by defining KPIs of interest and then recurringly obtaining data on those KPIs.
At 1210, a remedial operation is initiated based on the monitoring. Remedial operations may include operations such as migration, resource allocation or deallocation, restarting or suspending a process, container, pod, or microservice, or the like.
In existing orchestrated, distributed computing environments, hundreds or even thousands of microservices may be deployed. Software-based or software-implemented operators are used to manage the distributed resources and executable components. An operator may configure hardware or software based on SLO/SLA policies. Most operators that manage hardware are owned or provided by the platform provider, but some operators that manage software may be provided by the customer. In general, operators are software constructs that are used to setup, monitor, configure, and perform operations at the datacenter level. A software-implemented operator may be deployed with a container or workload. In Kubernetes, an operator is an application-specific controller that can help you package, deploy, and manage a Kubernetes application. Operators extend the functionality of Kubernetes and provide automation. Software agents, facilities, and utilities may provide the same or similar service as operators.
Various mechanisms are used to attest and validate platforms. For example, a Trusted Platform Module (TPM) may be used to provide hardware-based, security-related functions. A TPM chip is a secure crypto-processor that provides physical security mechanism to make it tamper-resistant. TPM is often used to ensure the integrity of a hardware platform. In contrast, operators do not have a similar attestation mechanism to prove its trustworthiness. Further, operators do not have mechanisms to attest to reputation or reliability.
While cloud service providers largely control a datacenter, suppliers may retain some control over products they produce, sell, or support such as firmware or microcode updates. Each component or sub-component (recursively) may have one or more suppliers that retain control over firmware updates, settings, and provisioning interfaces. Hence, datacenter and management outsource service providers may not fully delegate management control.
Consequently, the hardware mechanisms (which include roots of trust) may have identities, secrets, and trust anchors in secure storage for each supply chain entity that retains control as well as for any “owner” entity or management service provider (MSP) that may be delegated control as part of lifecycle management and onboarding.
TPM, DICE (Device Identity Composition Engine), DICE Protection Module (DPM), and RPMB (Replay Protected Memory Block) are examples of root-of-trust technologies. These may be used to establish identities, reputations, attestations, and trust policies for each ecosystem stakeholder/owner during manufacturing, supply and through deployment.
The orchestrator system may further verify attestation, ownership and identity values against a policy that establishes rightful suppliers/ownership/service provider relationships. Subsequent orchestration actions may depend on this policy being satisfied first.
The present systems and methods provide a mechanism for a customer supply a risk tolerance SLA. The risk tolerance SLA may be used for several purposes. For example, during execution, operators may be evaluated for reliability or reputation and only those that satisfy a threshold rating in view of the risk tolerance SLA may be used. As another example, platform capabilities may be evaluated in view of the risk tolerance SLA and certain platforms or configurations may be filtered or selected based on the risk tolerance indicated in the SLA or policies derived therefrom. Additionally, in implementations where tasks are able to share or borrow resources, operator reputation may be used as part of the inter-task negotiations.
Data center architectures are evolving in order to be capable to allow rapid provisioning of nodes, autonomous life cycle management of services, and seamless updates when needed. Ecosystem providers are moving towards developing methods that use constructs, such as operators in order to implement the automation, robustness, and life cycle management of the entire data center and related services. These operators can be developed on top of multiple types of technologies and languages. Operators may be involved in different phases of the system configuration and application deployment, and life cycle management depending on the nature of the operators.
The systems and methods described here provide orchestration and infrastructure to monitor how operators behave in different circumstances. For instance, operators may fail, change behavior of systems or applications in a non-desired way, etc. due to different loads, different type of services, etc. Over time the system can correlate operator behavior and system conditions before applying an operation a customer requests. Given a current situation, the system may check the likelihood of having a problem with one operator and decide to choose another operator.
The operator selection controller 1302 accesses a risk tolerance rating, which may be derived from or extracted from a risk tolerance intent SLA. The risk tolerance rating may be a numerical value in a certain range (e.g., from 0 to 7), for example. A rule database is used to determine a minimum reputation score that satisfies the risk tolerance rating and the risk tolerance intent SLA.
The set of operators is then identified from the minimum reputation score. Operators that are available to the platform may be stored in an operator database 1312. The operator database 1312 may be used to store operator information, such as an operator identifier, the type of operator, performance metrics, or other related data of the operator. The operator reputation scores may be stored in the operator database 1312. The operator selection controller 1302 selects operators that qualify based on the minimum reputation score.
The database of operator reputation scores may also be stored in a blockchain 1308 to ensure that the reputation scores are securely stored. A blockchain management logic 1310 may access the blockchain 1308 to retrieve current reputations for operators, update reputations, and the like. The operator selection controller 1302 and blockchain management logic 1310 effectively match a consumer's risk tolerance to operators with a sufficient reputation score.
A blockchain or distributed ledger technology (DLT) or distributed consensus algorithm (e.g., byzantine generals) may be used among multiple entities that evaluate reputation data and set reputation scores. The consensus algorithms ensure there is a diverse community that agrees such that a minority consensus cannot be applied to influence reputation.
A blockchain or DLT may additionally be used to maintain logged or historical information that cannot easily be rewritten to falsify facts. A consensuses algorithm may also establish the veracity of the fact as well as to prevent loss or modification to a recorded fact.
Reputation can be composed of many factors including, but not limited to reliability (e.g., number of subsequent infrastructure faults after deployment), service mesh disruption telemetry, application SLA violations, number of life cycle events triggered, and number of security violations. These types of feedback can be provided by infrastructure telemetry, orchestration telemetry, or service telemetry.
The operator selection controller 1302 provides monitor rules to an operator monitor logic 1304. The operator monitor logic 1304 is used to score the impact of the operator by monitoring infrastructure telemetry and telemetry from other sources. The monitor rules may include parameters, which may include the start time of the operator, resources modified, infrastructure endpoints, evaluation window, telemetry collectors, and guardrail values for telemetry KPIs.
An operator analytics logic 1306 may collect alerts from the operator monitor logic 1304 for excursions that occur during the evaluation window. These excursions are used to score the operator. The score is fed back to the operator database 1312 and used for future selections by the operator selection controller 1302. The operator can be scored on multiple dimensions including security, reliability, privacy, isolation/multi-tenancy, or cost.
At 1408, reputations of the operator are filtered based on the reputation selector function, if provided in the command. The reputation of the operator may be affected by various factors including, but not limited to historical operation of the operator, source of the operator, target platform that the operator will execute on, etc.
At 1410, it is determined whether the reputation score is higher than the minimum reputation score in the request command. If it is not, then the method 1400 terminates and the request to execute the operator is rejected (operation 1412). When the execution is rejected, the host will not configure the device. A notification may be generated to notify the device owner, notify the operator owner, or the like. There may be some situations where the host may take a calculated risk and execute operator if the device is one with a lower value (e.g., not a mission critical device).
If the reputation score is higher than the minimum reputation score, then the method 1400 continues. At 1414, the operator is executed. The execution result is checked (operation 1416) and results are sent to a blockchain (operation 1418). The execution results may be attested by the platform upon which the operation executed. At 1420, a new reputation score is computed based on the historical score. The new reputation score is written into the blockchain (operation 1422).
In response to determining that the software-implemented operator requires the trust evaluation, performing the following operations.
At 1504, the method 1500 includes obtaining a reputation score of the software-implemented operator. In an embodiment, the reputation score is based on a number of subsequent faults after deployment of the software-implemented operator. In a related embodiment, the reputation score is based on telemetry indicating service mesh disruption caused by the software-implemented operator. In a related embodiment, the reputation score is based on application service level agreement (SLA) violations caused by the software-implemented operator. In a related embodiment, the reputation score is based on a number of life cycle events caused by the software-implemented operator. In a related embodiment, the reputation score is based on a number of security violations caused by the software-implemented operator.
At 1506, the method 1500 includes determining a minimum reputation score from the risk tolerance intent.
At 1508, the method 1500 includes comparing the reputation score of the software-implemented operator to the minimum reputation score.
At 1510, the method 1500 includes rejecting or permitting execution of the software-implemented operator based on the comparison.
In an embodiment, the method 1500 includes monitoring the execution of the software-implemented operator and checking an execution result against an expected execution result. In a further embodiment, the execution result is attested by the compute node. In a related embodiment, the method 1500 includes storing the execution result in a blockchain. In a further embodiment, the method 1500 includes computing a revised reputation score based on the execution result and storing the revised reputation score in the blockchain.
In an embodiment, computing the revised reputation score involves evaluating historical operation of the software-implemented operator, a source of the software-implemented operator, or the compute node that the software-implemented operator executed on.
In an embodiment, the method 1500 includes determining based on analytics whether a task is threatened by a security risk and initiating a responsive action that is consistent with the risk tolerance intent.
Existing deployment models use either cloud native deployments inside of a virtual machine (VM) or pure cloud native approaches. In the former model, tenants are isolated in different VMs and pods from services share the same virtual space. In the latter model, tenants share the same environment. However, in either case, the entities are separated—there are either multiple VMs with multiple pods per VM, or there are multiple pods. These entities (VMs or pods) cannot communicate with each other to offer to share resources.
Systems and methods described herein provide a resource request broker that is integrated into the orchestration system. The resource request broker may maintain a message-based system, independent of the application KPI (which is monitored by the application itself). The broker may have a local agent on the platform that can make local decisions based on sharing policies. Sharing policies may be set at the host level, the cluster level, or other arrangements.
Individual pods can sign up for static contracts, “guaranteed resource contracts,” equal share contracts, or dynamic (bidding) contracts. Competing requests are arbitrated by the local broker agent, which can operate a bid-offer scheme. For instance, an entity may indicate “I'm willing to give up my resource X for your resource Y for period N.” As another example, a pod may indicate “I can reduce or release some compute capacity for some time given I'm not doing anything important or I can reduce my SLA.” This provides a mechanism for pods or VMs to know that another peer is willing to share some resource and then borrow it during some time to absorb a given load of peak.
This may also work the other way around where one VM or pod needs more horsepower for a peak. The pod or VM may request “Is there any peer from which I can borrow more compute for some time?” In addition, the request may indicate a willingness to pay for such service. “Hey, can anyone lend me some compute? I'm willing to pay X.” When agreements are met, they can be provided to the orchestration entity to close the contracts.
For an application 1602A, 1602B, when demand exceeds a predefined threshold for a resource that the application 1602A, 1602B is using, a resource demand request is generated. The resource demand request may be time constrained (e.g., have an effective time period to satisfy the request). Conversely, when resource utilization drops below a threshold over a predefined range, while the corresponding SLA is in a specified allowable range, a resource available notification is generated. The resource available notification may include a tolerance to disruption of the SLA when reassigning resources (e.g., a maximum allotment of re-allocatable resources so the application does not violate an SLA).
The resource demand requests and resource available notifications are transmitted to an SLA arbiter 1610. The SLA arbiter 1610 aggregates requests and notifications and evaluates which requests/notifications can be met based on rules. When processing the requests and notifications, the SLA arbiter 1610 may use resource usage telemetry from an SLA resource monitor 1618, reputations of the entities, operators associated with the entities, etc. to determine whether to attempt to satisfy a request or process a notification. After a resource reallocation has been approved, a resource operator 1612 adjusts allocation of specific resources at run-time. The resource operator 1612 maintains a record of the change and reverts it after the time period indicated in the request or notification.
If the SLA arbiter 1610 determines to move an application to a different platform, the SLA arbiter 1610 sends a command to the orchestration scheduler 1614 via the resource operator 1612 to reposition the application. If the SLA arbiter 1610 determines to reallocate resources on the shared platform, a command is sent to a resource control agent 1616 via the resource operator 1612 to adjust resource allocations.
At 1704, in response to the communication being the request for additional resources, it is determined whether resources are available at the platform, and if so, the resources are allocated to the application.
At 1706, in response to the communication being the request for additional resources, it is determined whether resources are available at the platform, and if not, a migration to move the application to a different platform is initiated.
At 1708, in response to the communication being a notification of available resources, it is determined whether a second application executing on the platform has requested resources that match the available resources, and if so, a loan of the available resources is facilitated to the second application.
In general, risk tolerance encompasses both a facet of trust and also a facet of security. Trust is reflected using reputation. In existing systems, security is handled as a binary decision—does the client require a secure environment or not? In a multi-tenant deployment environment, the question of security becomes more complicated. For instance, assuming a workload runs on a single-tenant XPU, it may be fine for the service owner to allow a co-located tenant as long as the transport-level security is enabled. However, a more security-sensitive tenant may require an exclusive environment (i.e., that the environment is not shared with any other tenant) or maybe only shared with trusted neighbors. This type of security tuning is not possible using a binary decision maker. Instead, what is needed is a more quantitative security mechanism to provide a more efficient and cost-effective deployment configuration.
A portion of a risk tolerance intent may be a security intent. The security intent may be an expression by a tenant or tenant owner of security features or objectives that it requires. The tenant or customer may express the security intent for a workload as privacy and security risk statements. For instance, the expression may be in the form of:
The systems and methods described herein enable the definition of a set of security-related intent declarations. Customers or service owners are able to define risk and threat profiles that they require to be satisfied during execution of their services or applications. A threat is an event that may happen. A risk is the probability of one or more threats occurring. Risks may be bracketed into a high, medium, and low probability zone. Each probability zone may be assigned a numerical value (e.g., a number in the range of [0,1] with the highest risk being a 1.0 and the lowest risk being 0).
Various threats may be accounted for, including but not limited to misconfigurations mitigated with policy enforcement, side channel attack mitigated through automatic anti-affinity settings, compromised credentials or duplicate shared credentials mitigated through password scanning, zero-day vulnerabilities mitigated via scan for known vulnerability fixes applied, or spy/ransomware attacks mitigated through software stack attestation.
Because services are setup differently, risks to them also needs to be addressed differently. For example, a service consisting of two components that rarely communicate needs to be addressed differently from a service consisting of many components in constant communication. The systems and methods described here add a set of components to the orchestration control plane to monitor, analyze, and assess risk. Another new component is able to translate the assessment into a set of policies, resource allocations, etc.
The API 1906 is used to configure the security manager 1900. The API 1906 may also be used to inspect the security level, risk factors, or threats for apps running in each context that is monitored by the security manager 1900.
The security manager 1900 includes a risk assessment engine 1908, a threat assessment engine 1910, a risk mitigation engine 1912, and a threat mitigation engine 1914. The security manager 1900 uses a policy database 1916 to determine responsive actions to detected risks and threats.
The threat assessment engine 1910 processes real-time security analytics from the monitoring facilities 1904 to identify potential threats. Similarly, the risk assessment engine 1908 processes real-time security analytics from the monitoring facilities 1904 to identify the probability of potential threats (i.e., risks).
The main tasks of the risk assessment engine 1908 and the threat assessment engine 1910 is to, for a given workload, determine the best possible security mechanism to fulfill the risk intent and threat mitigation as declared by the service owner. This will be different for each service, customer, and hardware platform at play.
For instance, if a workload accesses a dataset that is highly confidential and geo-fenced, the workload may be placed in a cluster that is consistent with the geo-fencing of the dataset. Additionally, the dataset may be encrypted, using TLS for transport, etc. to satisfy the threat mitigation policies. The initial security configuration may be annotated, for example, on deployment/pod spec level in a Kubernetes cluster. At runtime, the initial assessment and the derived security mechanisms are monitored and adapted if needed by the risk mitigation engine 1912 or the threat mitigation engine 1914. Depending on the execution state, the risk mitigation engine 1912 or the threat mitigation engine 1914 may manipulate the settings and polices in-place. The risk mitigation engine 1912 or the threat mitigation engine 1914 may proactively alter settings or policies if it results in better efficiency. Mechanisms may be adjusted during run time, for example, as some subtasks complete, other subtasks may be executing with a different security policy. Further, if the tenant workload is behaving badly, security policies may become stricter or the risk mitigation engine 1912 or the threat mitigation engine 1914 may move the workload to another platform.
The following is a list of security mechanisms that may be employed by the risk mitigation engine 1912 or the threat mitigation engine 1914. Various isolation mechanisms may be used, including language-level isolation, process-level isolation, virtualization, the use of Trusted Execution Environments (e.g., SGX, TDX, ARM Trustzone, Realms, etc.), or execution on dedicated core. Various communication security mechanisms may be used including use of IPSec, TLS, secure transports, link-level security, cryptographic algorithms, post-quantum crypto, or varying cipher strengths. When data is at rest, storage encryption may be used. For dealing with affinity or anti-affinity settings, the orchestration may avoid co-location of certain tasks. Other security mechanisms or factors include image (e.g., container image) security, role-based access control enforcement, root access requirements, and post-deployment clean up (e.g., writing random data to scratchpads/drives).
Security policies may include spatial deployment policies and temporal deployment policies. Spatial deployment policies include aspects like geo-locking locations of deployment, workload exclusions where one workload cannot be co-located with another workload having specific characteristics, and requiring random placement of workloads in a resource pool. Temporal deployment policies include aspects like requiring that a workload must be migrated to a random location in the resource pool at a specified interval (e.g., to employ a moving target defense).
The security manager 1900 provides an intent-driven orchestration function to make resource allocation decisions based on the tenant intent and real-time security analytics. The risk assessment engine 1908 and threat assessment engine 1910 provide security analytics functions. The risk assessment engine 1908 and threat assessment engine 1910 consume the resource pool telemetry. For example, resource pool telemetry is processed to assess whether workloads exhibit malicious behaviors using performance counters. Workloads that are flagged can be terminated or migrated to an execution environment with stronger security guarantees, e.g. from an environment using process isolation to a virtualized environment. The risk assessment engine 1908 and threat assessment engine 1910 may also consider attestation measurements. Attestation mechanisms enable the analytics function to establish trust in the resource pool in terms of its constituents, e.g., execution environment integrity, platform integrity, and platform identity.
The attestation may provide a confidence score or reputation score to establish how well certain resources, or point-of-presence perform. This can be a key input for the mitigation and remediation provided by the risk mitigation engine 1912 or the threat mitigation engine 1914. A blockchain-based determination may be used when there is no other way for trust/credibility to be established through a third party. Initially, a platform or node may not have any history of its performance A ramp up stage is when a node builds a certain record of trust. The node may take any workload from clients that are willing to use a new node. Third party brokering services may provide test-and-attestation services for such incipient nodes, to jump start their client streams (to deal with startup inertia). Clients with less sensitive data or lower levels of quality of service may use the node because it presents a lower overall risk. Over time, the node is able to build a reputation for itself and progressively attract other clients who are willing to trust it. After interactions with the node, each of the node and the customer (i.e., application, virtual machine, container, etc.) may add a reputation adjustment to the node's reputation score in a blockchain. In the steady state, a blockchain provides an entire record of the trust met or trust violation history of a node. There is onus on both the consumer and provider to enter their transactions into the blockchain so that there is a public, immutable record of what was negotiated, what was violated, etc. The reputation score or profile is useful for setups crossing multiple points-of-presence (cloud, edge, etc.).
At 2004, analytics of the task during execution are received. The analytics may include environmental settings, contextual settings, operational metrics, and the like.
At 2006, the analytics are evaluated to identify whether a security threat exists.
At 2008, in response to identifying that the security threat exists, access a policy database to determine a responsive action. The responsive action may include various operations, such as adding or removing resource allocations to the task, initiating language-level isolation of the task, initiating process-level isolation of the task, initiating virtualization of the task, executing the task in a trusted execution environment, or executing the task on a dedicated core. The responsive action may include implementing a communication security mechanism to secure communications by the task. The responsive action may include encrypting data that is accessed by the task.
At 2010, the responsive action is initiated. This may be conducted by way of a resource allocation operator, an orchestrator, or the like.
Orchestrator or orchestration systems described in this document may be implemented in an appliance, endpoint component, device, client device, computing device, a node, a server, a standalone node that is sold separately and integrated into a different system, etc. A system architect may implement an orchestration system in any installation.
Embodiments may be implemented in one or a combination of hardware, firmware, and software. Embodiments may also be implemented as instructions stored on a machine-readable storage device, which may be read and executed by at least one processor to perform the operations described herein. A machine-readable storage device may include any non-transitory mechanism for storing information in a form readable by a machine (e.g., a computer). For example, a machine-readable storage device may include read-only memory (ROM), random-access memory (RAM), magnetic disk storage media, optical storage media, flash-memory devices, and other storage devices and media.
Examples, as described herein, may include, or may operate on, logic or a number of components, such as modules, intellectual property (IP) blocks or cores, engines, or mechanisms. Such logic or components may be hardware, software-configured hardware, or firmware communicatively coupled to one or more processors in order to carry out the operations described herein. Logic or components may be hardware modules (e.g., IP block), and as such may be considered tangible entities capable of performing specified operations and may be configured or arranged in a certain manner In an example, circuits may be arranged (e.g., internally or with respect to external entities such as other circuits) in a specified manner as an IP block, IP core, system-on-chip (SoC), or the like.
In an example, the whole or part of one or more computer systems (e.g., a standalone, client or server computer system) or one or more hardware processors may be configured by firmware or software (e.g., instructions, an application portion, or an application) as a module that operates to perform specified operations. In an example, the software may reside on a machine-readable medium. In an example, the software, when executed by the underlying hardware of the module, causes the hardware to perform the specified operations. Accordingly, the term hardware module is understood to encompass a tangible entity, be that an entity that is physically constructed, specifically configured (e.g., hardwired), or temporarily (e.g., transitorily) configured (e.g., programmed) to operate in a specified manner or to perform part or all of any operation described herein.
Considering examples in which modules are temporarily configured, each of the modules need not be instantiated at any one moment in time. For example, where the modules comprise a general-purpose hardware processor configured using software; the general-purpose hardware processor may be configured as respective different modules at different times. Software may accordingly configure a hardware processor, for example, to constitute a particular module at one instance of time and to constitute a different module at a different instance of time. Modules may also be software or firmware modules, which operate to perform the methodologies described herein.
An IP block (also referred to as an IP core) is a reusable unit of logic, cell, or integrated circuit. An IP block may be used as a part of a field programmable gate array (FPGA), application-specific integrated circuit (ASIC), programmable logic device (PLD), system on a chip (SoC), or the like. It may be configured for a particular purpose, such as digital signal processing or image processing. Example IP cores include central processing unit (CPU) cores, integrated graphics, security, input/output (I/O) control, system agent, graphics processing unit (GPU), artificial intelligence, neural processors, image processing unit, communication interfaces, memory controller, peripheral device control, platform controller hub, or the like.
Example 1 is a system for implementing intent-driven security mechanisms, comprising: a processor; and memory to store instructions, which when executed by the processor, cause the system to: determine, based on a risk tolerance intent related to execution of an application on a compute node, whether execution of a software-implemented operator requires a trust evaluation; and in response to determining that the software-implemented operator requires the trust evaluation: obtain a reputation score of the software-implemented operator; determine a minimum reputation score from the risk tolerance intent; compare the reputation score of the software-implemented operator to the minimum reputation score; and reject or permit execution of the software-implemented operator based on the comparison.
In Example 2, the subject matter of Example 1 includes, wherein the reputation score is based on a number of subsequent faults after deployment of the software-implemented operator.
In Example 3, the subject matter of Examples 1-2 includes, wherein the reputation score is based on telemetry indicating service mesh disruption caused by to the software-implemented operator.
In Example 4, the subject matter of Examples 1-3 includes, wherein the reputation score is based on application service level agreement (SLA) violations caused by the software-implemented operator.
In Example 5, the subject matter of Examples 1-4 includes, wherein the reputation score is based on a number of life cycle events caused by the software-implemented operator.
In Example 6, the subject matter of Examples 1-5 includes, wherein the reputation score is based on a number of security violations caused by the software-implemented operator.
In Example 7, the subject matter of Examples 1-6 includes, wherein the system is to: monitor the execution of the software-implemented operator; and check an execution result against an expected execution result.
In Example 8, the subject matter of Example 7 includes, wherein the execution result is attested by the compute node.
In Example 9, the subject matter of Examples 7-8 includes, wherein the system is to store the execution result in a blockchain.
In Example 10, the subject matter of Example 9 includes, wherein the system is to: compute a revised reputation score based on the execution result; and store the revised reputation score in the blockchain.
In Example 11, the subject matter of Example 10 includes, wherein to compute the revised reputation score, the system is to evaluate historical operation of the software-implemented operator, a source of the software-implemented operator, or the compute node that the software-implemented operator executed on.
In Example 12, the subject matter of Examples 1-11 includes, where the system is to: determine based on analytics whether a task is threatened by a security risk; and initiate a responsive action that is consistent with the risk tolerance intent.
Example 13 is a method for implementing intent-driven security mechanisms, comprising: determining, based on a risk tolerance intent related to execution of an application on a compute node, whether execution of a software-implemented operator requires a trust evaluation; and in response to determining that the software-implemented operator requires the trust evaluation: obtaining a reputation score of the software-implemented operator; determining a minimum reputation score from the risk tolerance intent; comparing the reputation score of the software-implemented operator to the minimum reputation score; and rejecting or permitting execution of the software-implemented operator based on the comparison.
In Example 14, the subject matter of Example 13 includes, wherein the reputation score is based on a number of subsequent faults after deployment of the software-implemented operator.
In Example 15, the subject matter of Examples 13-14 includes, wherein the reputation score is based on telemetry indicating service mesh disruption caused by to the software-implemented operator.
In Example 16, the subject matter of Examples 13-15 includes, wherein the reputation score is based on application service level agreement (SLA) violations caused by the software-implemented operator.
In Example 17, the subject matter of Examples 13-16 includes, wherein the reputation score is based on a number of life cycle events caused by the software-implemented operator.
In Example 18, the subject matter of Examples 13-17 includes, wherein the reputation score is based on a number of security violations caused by the software-implemented operator.
In Example 19, the subject matter of Examples 13-18 includes, monitoring the execution of the software-implemented operator; and checking an execution result against an expected execution result.
In Example 20, the subject matter of Example 19 includes, wherein the execution result is attested by the compute node.
In Example 21, the subject matter of Examples 19-20 includes, storing the execution result in a blockchain.
In Example 22, the subject matter of Example 21 includes, computing a revised reputation score based on the execution result; and storing the revised reputation score in the blockchain.
In Example 23, the subject matter of Example 22 includes, wherein computing the revised reputation score includes evaluating historical operation of the software-implemented operator, a source of the software-implemented operator, or the compute node that the software-implemented operator executed on.
In Example 24, the subject matter of Examples 13-23 includes, determining based on analytics whether a task is threatened by a security risk; and initiating a responsive action that is consistent with the risk tolerance intent.
Example 25 is at least one machine-readable medium including instructions, which when executed by a machine, cause the machine to perform operations of any of the methods of Examples 13-24.
Example 26 is an apparatus comprising means for performing any of the methods of Examples 13-24.
Example 27 is at least one machine-readable medium including instructions for implementing intent-driven security mechanisms, which when executed by a machine, cause the machine to perform operations comprising: determining, based on a risk tolerance intent related to execution of an application on a compute node, whether execution of a software-implemented operator requires a trust evaluation; and in response to determining that the software-implemented operator requires the trust evaluation: obtaining a reputation score of the software-implemented operator; determining a minimum reputation score from the risk tolerance intent; comparing the reputation score of the software-implemented operator to the minimum reputation score; and rejecting or permitting execution of the software-implemented operator based on the comparison.
In Example 28, the subject matter of Example 27 includes, wherein the reputation score is based on a number of subsequent faults after deployment of the software-implemented operator.
In Example 29, the subject matter of Examples 27-28 includes, wherein the reputation score is based on telemetry indicating service mesh disruption caused by to the software-implemented operator.
In Example 30, the subject matter of Examples 27-29 includes, wherein the reputation score is based on application service level agreement (SLA) violations caused by the software-implemented operator.
In Example 31, the subject matter of Examples 27-30 includes, wherein the reputation score is based on a number of life cycle events caused by the software-implemented operator.
In Example 32, the subject matter of Examples 27-31 includes, wherein the reputation score is based on a number of security violations caused by the software-implemented operator.
In Example 33, the subject matter of Examples 27-32 includes, instructions for: monitoring the execution of the software-implemented operator; and checking an execution result against an expected execution result.
In Example 34, the subject matter of Example 33 includes, wherein the execution result is attested by the compute node.
In Example 35, the subject matter of Examples 33-34 includes, storing the execution result in a blockchain.
In Example 36, the subject matter of Example 35 includes, instructions for: computing a revised reputation score based on the execution result; and storing the revised reputation score in the blockchain.
In Example 37, the subject matter of Example 36 includes, wherein computing the revised reputation score includes evaluating historical operation of the software-implemented operator, a source of the software-implemented operator, or the compute node that the software-implemented operator executed on.
In Example 38, the subject matter of Examples 27-37 includes, instructions for: determining based on analytics whether a task is threatened by a security risk; and initiating a responsive action that is consistent with the risk tolerance intent.
Example 39 is an apparatus for implementing intent-driven security mechanisms, comprising: means for determining, based on a risk tolerance intent related to execution of an application on a compute node, whether execution of a software-implemented operator requires a trust evaluation; and in response to determining that the software-implemented operator requires the trust evaluation: means for obtaining a reputation score of the software-implemented operator; means for determining a minimum reputation score from the risk tolerance intent; means for comparing the reputation score of the software-implemented operator to the minimum reputation score; and means for rejecting or permitting execution of the software-implemented operator based on the comparison.
In Example 40, the subject matter of Example 39 includes, wherein the reputation score is based on a number of subsequent faults after deployment of the software-implemented operator.
In Example 41, the subject matter of Examples 39-40 includes, wherein the reputation score is based on telemetry indicating service mesh disruption caused by to the software-implemented operator.
In Example 42, the subject matter of Examples 39-41 includes, wherein the reputation score is based on application service level agreement (SLA) violations caused by the software-implemented operator.
In Example 43, the subject matter of Examples 39-42 includes, wherein the reputation score is based on a number of life cycle events caused by the software-implemented operator.
In Example 44, the subject matter of Examples 39-43 includes, wherein the reputation score is based on a number of security violations caused by the software-implemented operator.
In Example 45, the subject matter of Examples 39-44 includes, means for monitoring the execution of the software-implemented operator; and means for checking an execution result against an expected execution result.
In Example 46, the subject matter of Example 45 includes, wherein the execution result is attested by the compute node.
In Example 47, the subject matter of Examples 45-46 includes, storing the execution result in a blockchain.
In Example 48, the subject matter of Example 47 includes, means for computing a revised reputation score based on the execution result; and means for storing the revised reputation score in the blockchain.
In Example 49, the subject matter of Example 48 includes, wherein computing the revised reputation score includes evaluating historical operation of the software-implemented operator, a source of the software-implemented operator, or the compute node that the software-implemented operator executed on.
In Example 50, the subject matter of Examples 39-49 includes, means for determining based on analytics whether a task is threatened by a security risk; and means for initiating a responsive action that is consistent with the risk tolerance intent.
Example 51 is at least one machine-readable medium including instructions that, when executed by processing circuitry, cause the processing circuitry to perform operations to implement of any of Examples 1-50.
Example 52 is an apparatus comprising means to implement of any of Examples 1-50.
Example 53 is a system to implement of any of Examples 1-50.
Example 54 is a method to implement of any of Examples 1-50.
The above detailed description includes references to the accompanying drawings, which form a part of the detailed description. The drawings show, by way of illustration, specific embodiments that may be practiced. These embodiments are also referred to herein as “examples.” Such examples may include elements in addition to those shown or described. However, also contemplated are examples that include the elements shown or described. Moreover, also contemplated are examples using any combination or permutation of those elements shown or described (or one or more aspects thereof), either with respect to a particular example (or one or more aspects thereof), or with respect to other examples (or one or more aspects thereof) shown or described herein.
Publications, patents, and patent documents referred to in this document are incorporated by reference herein in their entirety, as though individually incorporated by reference. In the event of inconsistent usages between this document and those documents so incorporated by reference, the usage in the incorporated reference(s) are supplementary to that of this document; for irreconcilable inconsistencies, the usage in this document controls.
In this document, the terms “a” or “an” are used, as is common in patent documents, to include one or more than one, independent of any other instances or usages of “at least one” or “one or more.” In this document, the term “or” is used to refer to a nonexclusive or, such that “A or B” includes “A but not B,” “B but not A,” and “A and B,” unless otherwise indicated. In the appended claims, the terms “including” and “in which” are used as the plain-English equivalents of the respective terms “comprising” and “wherein.” Also, in the following claims, the terms “including” and “comprising” are open-ended, that is, a system, device, article, or process that includes elements in addition to those listed after such a term in a claim are still deemed to fall within the scope of that claim. Moreover, in the following claims, the terms “first,” “second,” and “third,” etc. are used merely as labels, and are not intended to suggest a numerical order for their objects.
The above description is intended to be illustrative, and not restrictive. For example, the above-described examples (or one or more aspects thereof) may be used in combination with others. Other embodiments may be used, such as by one of ordinary skill in the art upon reviewing the above description. The Abstract is to allow the reader to quickly ascertain the nature of the technical disclosure. It is submitted with the understanding that it will not be used to interpret or limit the scope or meaning of the claims. Also, in the above Detailed Description, various features may be grouped together to streamline the disclosure. However, the claims may not set forth every feature disclosed herein as embodiments may feature a subset of said features. Further, embodiments may include fewer features than those disclosed in a particular example. Thus, the following claims are hereby incorporated into the Detailed Description, with a claim standing on its own as a separate embodiment. The scope of the embodiments disclosed herein is to be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled.
This application claims priority to U.S. Provisional Patent Application No. 63/280,001, filed Nov. 16, 2021, which is incorporated herein by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 63280001 | Nov 2021 | US |
