Patent Application
20230195601
Embodiments relate generally to data processing and more particularly to synthetic data generation for enhanced microservice debugging in microservices architectures.
Datacenters often leverage a microservice architecture to provide for network infrastructure services. A microservice architecture can arrange an application as a collection of loosely-coupled microservices. Microservices can refer to processes that communicate over a network to fulfill a goal using technology-agnostic protocols. In some cases, the microservices may be deployed using a container orchestration platform providing containerized workloads and/or services. The container orchestration platforms may utilize a service mesh to manage the high volume of network-based inter-process communication among the microservices. The service mesh is a dedicated software infrastructure layer for the microservices that includes elements to enable the communication among the microservices to be fast, reliable, and secure. The service mesh provides capabilities including service discovery, load balancing, encryption, observability, traceability, and authentication and authorization. The microservices deployment model provided by the service mesh is becoming increasingly elastic, providing flexibility to scale up and scale down microservices.
In a service mesh environment, a typical worker node in a compute cluster can handle hundreds of container workloads at the same time. These worker nodes may also have statically-attached specialized hardware accelerators optimized for compute intensive tasks. For instance, a class of hardware accelerators can be optimized to efficiently run cryptography and compression algorithms, or to run machine-learning acceleration algorithms. Such hardware accelerators may be provided as a form of disaggregated computing, where the workloads are distributed on disaggregated compute resources, such as CPUs, GPUs, and hardware accelerators (including field programmable gate arrays (FPGAs)), that are connected via a network instead of being on the same platform and connected via physical links such as peripheral component interconnect express (PCIe). Disaggregated computing enables improved resource utilization and lowers ownership costs by enabling more efficient use of available resources. Disaggregated computing also enables pooling a large number of hardware accelerators for large computation making the computation more efficient and better performing.
The microservices deployment model provided by the service mesh is becoming increasingly elastic, providing flexibility to scale up and scale down microservices. As the elasticity of deployment of microservices increases and as microservices architecture transitions to utilizing disaggregated computing resources, there can be microservices deployed for a service across many heterogeneous hardware devices. As such, it is becoming increasingly difficult to provide fine-grained tracing capabilities, including synthetic data generation to serve dynamic query generation and injection in such microservices architectures.
So that the manner in which the above recited features of the present embodiments can be understood in detail, a more particular description of the embodiments, briefly summarized above, may be had by reference to embodiments, some of which are illustrated in the appended drawings. It is to be noted, however, that the appended drawings illustrate typical embodiments and are therefore not to be considered limiting of its scope. The figures are not to scale. In general, the same reference numbers are used throughout the drawing(s) and accompanying written description to refer to the same or like parts.
Implementations herein describe synthetic data generation for enhanced microservice debugging in microservices architectures.
Cloud service providers (CSPs) are deploying solutions in datacenters where processing of a workload is distributed on various compute resources, such as central processing units (CPUs), graphics processing units (GPUs), and/or hardware accelerators (including, but not limited to, GPUs, field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), cryptographic accelerators, compression accelerators, and so on). Traditionally, these compute resources were running on the same platform and connected via physical communication links, such as peripheral component interconnect express (PCIe).
However, disaggregated computing is on the rise in datacenters. With disaggregated computing, CSPs are deploying solutions where processing of a workload is distributed on disaggregated compute resources, such as CPUs, GPUs, and hardware accelerators (including FPGAs, ASICs, etc.), that are connected via a network instead of being on the same platform and connected via physical links such as PCIe. Disaggregated computing enables improved resource utilization and lowers ownership costs by enabling more efficient use of available resources. Disaggregated computing also enables pooling a large number of hardware accelerators for large computation making the computation more efficient and better performing.
Hardware accelerators (also referred to herein as a hardware accelerator resources, hardware accelerator devices, accelerator resource, accelerator device, and/or extended resource) as discussed herein may refer to any of special-purpose central processing units (CPUs), graphics processing units (GPUs), general purpose GPUs (GPGPUs), field programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), inference accelerators, cryptographic accelerators, compression accelerators, other special-purpose hardware accelerators, and so on.
Moreover, the datacenters used by CSPs to deploy a service mesh often leverage a microservice architecture to provide for network infrastructure services of the service mesh. A microservice architecture can arrange an application as a collection of loosely-coupled microservices. The microservices may be the processes that communicate over a network to fulfill a goal using technology-agnostic protocols. In some cases, the microservices can be deployed using a container orchestration platform providing containerized workloads and/or services. In some examples, the service may be a large service comprising hundreds of microservices working in conjunction with each other or may be a modest individual service. A workload may refer to a resource running on the cloud consuming resources, such as computing power. In some embodiments, an application, service, or microservice may be referred to as a workload, which denotes the workload can be moved around between different cloud platforms or from on-premises to the cloud or vice-versa without any dependencies or hassle.
The container orchestration platforms may utilize a service mesh to manage the high volume of network-based inter-process communication among the microservices. The service mesh is a dedicated software infrastructure layer for the microservices that includes elements to enable the communication among the microservices to be fast, reliable, and secure. The service mesh provides capabilities including service discovery, load balancing, encryption, observability, traceability, and authentication and authorization.
As previously noted, the microservices deployment model provided by the service mesh is becoming increasingly elastic, providing flexibility to scale up and scale down microservices. As the elasticity of deployment of microservices increases and as microservices architecture transitions to utilizing disaggregated computing resources, there can be microservices deployed for a service across many heterogeneous hardware devices (e.g., intellectual property core or block (IPs), heterogeneous processing units (XPUs)). As such, it is becoming increasingly difficult to provide fine-grained tracing capabilities, including synthetic data generation to serve dynamic query generation and injection in such microservices architectures.
This especially apparent on conventional systems that lack the fine-granular tracing capability. In particular, such conventional systems lack a capability for synthetic data generation to be used in dynamic query ingestion for a given intricate interdependent set of microservices components. Furthermore, conventional systems lack a capability to discover an activation profile of the microservice hardware resources (e.g., XPUs and IPs) and associated software elements that can be used to provide for fine-grained debugging in today’s elastic and disaggregated microservices architectures.
Implementations herein address the above-noted technical drawbacks by providing for synthetic data generation for enhanced microservice debugging in microservices architectures. In implementations herein, techniques are provided for generating a synthetic dataset. Using information and recommendations from a discovery and recommender service, a synthetic data generator can provide for synthetic data generation for utilization with dynamic query generation and ingestion in enhanced debugging of a service. For example, based on discovered activation profiles of the service, configured polices for the service, an interdependency flow matrix of the service, and inputs on a recommended query to be exercised, the synthetic data generator can generate a synthetic dataset based on machine learning (ML)-based heuristics of past real infield data provided by the service. The ML-based heuristics and generated filters can be applied to an original dataset of the service to generate privacy-filtered and aggregated data of the synthetic dataset.
Implementations herein provide technical advantages over the conventional approaches discussed above. One technical advantage is that implementations provide for improved tracing, debugging, and performance evaluation utilizing the synthetic data generation to enhance dynamic query generation and query ingestion techniques described herein. The synthetic data generation techniques can be used to stress microservice component boundaries in a microservices architecture, while optionally providing a mirroring capability for failover and/or resiliency. This leads to improved performance in terms of processing speed and communication latency of the microservices architecture.
The datacenter of system 100 provides its hosted processing components 115, 125, 135, 145, 155, 162, 164, 166 with a variety of offloads using, for example, IPUs 105 that are directly attached to the respective host processing component. Although IPUs 105 are discussed for example purposes, other programmable network devices, such as DPUs or SmartNICs, may be used interchangeable for IPUs 105 herein. The offloads provided may be networking, storage, security, etc. This allows the processing components 115, 125, 135, 145, 155, 162, 164, 166 to run without a hypervisor, and provides CSPs the capability of renting out the host in a datacenter to their security-minded customers, or avoid cross-talk and other problems associated with multi-tenant hosts.
An IPU 105 can provide a role in datacenters by providing the datacenter operator, such as a Cloud Service Provider (CSP), a control point for security, acceleration, telemetry and service orchestration. IPU 105 architecture may build upon existing Smart Network Interface Card (SmartNIC) features and is a part of controlling security and data acceleration within and across distributed platforms. It is a secure domain controlled by CSPs for managing a platform, providing services to tenants, and securing access into the datacenter network. The IPU 105 increases the performance and predictability for distributed runtimes and enables scaling to multi-terabit throughputs by offloading host services, reliable transport, and optimizing data copies.
IPUs 105 have grown in complexity over the years, starting with foundational NICs, whose sole purpose was to get packets into the host and out of it. With the addition of networking software offload, the NICs evolved to become SmartNICs, that are capable of offloading functions, such as VSwitch, VIRTIO-Net, AVF, etc. Remote disaggregated storage architectures provide a further evolution, where compute and storage are not co-located anymore, but large compute clusters are connected to large storage clusters over the network. Increase in network speeds, and evolution of protocols made this a possibility. One of the advantages that remote disaggregated storage offers over direct attached storage is that compute and memory can be developed and updated at different cadences. The amount of memory that is attached to a compute node is not limited by physical addition or removal of hard-drives anymore, but can be hot-plugged as a PF to a PCIe Switch. Technologies such as Smart End Point enable IPUs to have firmware-controlled switches, along the PCIe Switch itself to not be limited by hardware implementations.
As discussed above, embodiments herein provide for synthetic data generation for enhanced microservice debugging in microservices architectures. In one implementation, datacenter system 100 includes one or more resources that can implement service management component 170 to provide telemetry targeted query injection for enhanced debugging in microservices architectures. For illustrative example purposes, service management component 170 is shown in the CPU 115 and GPU 135, respectively, of datacenter system 100. However, service management component 170 may operate in one or more of the various other disaggregated resources of datacenter system 100 in accordance with implementations herein. As such, the resources of datacenter system 100 may be in different platforms connected via a network (not shown) in the datacenter system 100. In some implementations, software and/or middleware can cause the resources of datacenter system 100 to logically appear to be in the same platform. Furthermore, transport protocols implemented in software and/or hardware (e.g., network interface cards (NICs)) can make the remote resources logically appear as if they are local resources as well.
Further details of the service management component 170 implementing the synthetic data generation for enhanced microservice debugging in microservices architectures is described below with respect to
A platform 202 may include platform resources 210 with one or more processing resources 212 (e.g., XPUs including CPUs, GPUs, FPGAs, ASICs, other hardware accelerators), memories 214 (which may include any number of different modules), chipsets 216, communication interface device(s) 218, and any other suitable hardware and/or software to execute a hypervisor 213 or other operating system capable of executing workloads associated with applications running on platform 202.
In some embodiments, a platform 202 may function as a host platform for one or more guest systems 222 that invoke these applications. Platform 202A may represent any suitable computing environment, such as a high-performance computing environment, a datacenter, a communications service provider infrastructure (e.g., one or more portions of an Evolved Packet Core), an in-memory computing environment, a computing system of a vehicle (e.g., an automobile or airplane), an Internet of Things (IoT) environment, an industrial control system, other computing environment, or combination thereof.
Each platform 202 may include platform resources 210. Platform resources 210 can include, among other logic enabling the functionality of platform 202, one or more processing resources 212 (such as CPUs, GPUs, FPGAs, other hardware accelerators, etc.), memory 214, one or more chipsets 216, and communication interfaces 228. Although three platforms are illustrated, computer platform 202A may be interconnected with any suitable number of platforms. In various embodiments, a platform 202 may reside on a circuit board that is installed in a chassis, rack, or other suitable structure that comprises multiple platforms coupled together through network 208 (which may comprise, e.g., a rack or backplane switch).
In the case of processing resources 212 comprising CPUs, the CPUs may each comprise any suitable number of processor cores and supporting logic (e.g., uncores). The cores may be coupled to each other, to memory 214, to at least one chipset 216, and / or to a communication interface device 218, through one or more controllers residing on the processing resource 212 (e.g., CPU) and/or chipset 216. In some embodiments, a processing resource 212 is embodied within a socket that is permanently or removably coupled to platform 202A. A platform 202 may include any suitable number of processing resources 212.
Memory 214 may comprise any form of volatile or nonvolatile memory including, without limitation, magnetic media (e.g., one or more tape drives), optical media, random access memory (RAM), read-only memory (ROM), flash memory, removable media, or any other suitable local or remote memory component or components. Memory 214 may be used for short, medium, and/or long term storage by platform 202A. Memory 214 may store any suitable data or information utilized by platform resources 210, including software embedded in a computer readable medium, and/or encoded logic incorporated in hardware or otherwise stored (e.g., firmware). Memory 214 may store data that is used by cores of processing resources 212. In some embodiments, memory 214 may also comprise storage for instructions that may be executed by the processing resources 212 (e.g., cores of CPUs) or other processing elements (e.g., logic resident on chipsets 216) to provide functionality associated with the management component 226 or other components of platform resources 210.
A platform 202 may also include one or more chipsets 216 comprising any suitable logic to support the operation of the processing resources 212. In various embodiments, chipset 216 may reside on the same die or package as a processing resource 212 or on one or more different dies or packages. Each chipset may support any suitable number of processing resources 212. A chipset 216 may also include one or more controllers to couple other components of platform resources 210 (e.g., communication interface 228 or memory 214) to one or more processing resources 212.
In the embodiment depicted, each chipset 216 also includes a management component 226. Management component 226 may include any suitable logic to support the operation of chipset 216. In a particular embodiment, a management component 226 can collect real-time telemetry data from the chipset 216, the processing resources 212, and/or memory 214 managed by the chipset 216, other components of platform resources 210, and/or various connections between components of platform resources 210.
Chipsets 216 also each include a communication interface 228. Communication interface 228 may be used for the communication of signaling and/or data between chipset 216 and one or more I/O devices, one or more networks 208, and/or one or more devices coupled to network 208 (e.g., datacenter management platform 206). For example, communication interface 228 may be used to send and receive network traffic such as data packets. In a particular embodiment, a communication interface 228 comprises one or more physical network interface controllers (NICs), also known as network interface cards or network adapters. A NIC may include electronic circuitry to communicate using any suitable physical layer and data link layer standard such as Ethernet (e.g., as defined by an IEEE 802.3 standard), FibreChannel, InfiniBand, Wi-Fi, or other suitable standard. A NIC may include one or more physical ports that may couple to a cable (e.g., an Ethernet cable). A NIC may enable communication between any suitable element of chipset 216 (e.g., management component 226) and another device coupled to network 208. In various embodiments, a NIC may be integrated with the chipset 216 (i.e., may be on the same integrated circuit or circuit board as the rest of the chipset logic) or may be on a different integrated circuit or circuit board that is electromechanically coupled to the chipset.
Platform resources 210 may include an additional communication interface 228. Similar to communication interface devices 218, communication interfaces 228 may be used for the communication of signaling and/or data between platform resources 210 and one or more networks 208 and one or more devices coupled to the network 208. For example, communication interface 228 may be used to send and receive network traffic such as data packets. In a particular embodiment, communication interfaces 228 comprise one or more physical NICs. These NICs may enable communication between any suitable element of platform resources 210 (e.g., processing resources 212 or memory 214) and another device coupled to network 208 (e.g., elements of other platforms or remote computing devices coupled to network 208 through one or more networks).
Platform resources 210 may receive and perform any suitable types of workloads. A workload may include any request to utilize one or more resources of platform resources 210, such as one or more cores or associated logic. For example , a workload may comprise a request to instantiate a software component, such as an I/O device driver 224 or guest system 222; a request to process a network packet received from a microservices container 232A, 232B (collectively referred to herein as microservice containers 232) or device external to platform 202A (such as a network node coupled to network 208); a request to execute a process or thread associated with a guest system 222, an application running on platform 202A, a hypervisor 213 or other operating system running on platform 202A; or other suitable processing request.
A microservice container 232 may emulate a computer system with its own dedicated hardware. A container 232 may refer to a standard unit of software that packages up code and all its dependencies, so the application runs quickly and reliably from one computing environment to another. A container image is a lightweight, standalone, executable package of software that includes components used to run an application: code, runtime, system tools, system libraries and settings. Containers 232 take advantage of a form of operating system (OS) virtualization in which features of the OS are leveraged to both isolate processes and control the amount of CPU, memory, and disk that those processes have access to.
When implementing containers 232, hypervisor 213 may also be referred to as a container runtime. Although implementations herein discuss virtualization of microservice functionality via containers, in some implementations, virtual machines may be hosted by hypervisor 213 and utilized to host microservices and/or other components of a service provided by an application.
A hypervisor 213 (also known as a virtual machine monitor (VMM)) may comprise logic to create and run guest systems 222. The hypervisor 213 may present guest operating systems run by virtual machines with a virtual operating platform (i.e., it appears to the virtual machines that they are running on separate physical nodes when they are actually consolidated onto a single hardware platform) and manage the execution of the guest operating systems by platform resources 210. Services of hypervisor 213 may be provided by virtualizing in software or through hardware-assisted resources that utilize minimal software intervention, or both. Multiple instances of a variety of guest operating systems may be managed by the hypervisor 213. Each platform 202 may have a separate instantiation of a hypervisor 213.
In implementations herein, the hypervisor 213 may also be implemented as a container runtime environment capable of building and containerizing applications.
Hypervisor 213 may be a native or bare-metal hypervisor that runs directly on platform resources 210 to control the platform logic and manage the guest operating systems. Alternatively, hypervisor 213 may be a hosted hypervisor that runs on a host operating system and abstracts the guest operating systems from the host operating system. Hypervisor 213 may include a virtual switch 238 that may provide virtual switching and/or routing functions to virtual machines of guest systems 222.
Virtual switch 238 may comprise a software element that is executed using components of platform resources 210. In various embodiments, hypervisor 213 may be in communication with any suitable entity (e.g., a SDN controller) which may cause hypervisor 213 to reconfigure the parameters of virtual switch 238 in response to changing conditions in platform 202 (e.g., the addition or deletion of microservice containers 232 or identification of optimizations that may be made to enhance performance of the platform).
The elements of platform resources 210 may be coupled together in any suitable manner. For example, a bus may couple any of the components together. A bus may include any known interconnect, such as a multi-drop bus, a mesh interconnect, a ring interconnect, a point-to-point interconnect, a serial interconnect, a parallel bus, a coherent (e.g., cache coherent) bus, a layered protocol architecture, a differential bus, or a Gunning transceiver logic (GTL) bus, to name a few examples.
Elements of the computer platform 202A may be coupled together in any suitable manner such as through one or more networks 208. A network 208 may be any suitable network or combination of one or more networks operating using one or more suitable networking protocols. A network may represent a series of nodes, points, and interconnected communication paths for receiving and transmitting packets of information that propagate through a communication system. For example, a network may include one or more firewalls, routers, switches, security appliances, antivirus servers, or other useful network devices.
In implementations herein, one or more of processing resources 212 and/or microservice containers 232 may provide a service management component (not shown), such as service management component 170 described with respect to
In one implementation, service platform 300 may host a service implemented with one or more microservice containers 320A, 320B (collectively referred to herein as microservice container 320). Microservice containers 320 may be the same as microservice containers 232 described with respect to
Service platform 300 may function as a host platform for a service, implementing deployed microservices of the service as one or more microservice containers 320 that invoke functionalities of the service. Service platform 300 may represent any suitable computing environment, such as a high-performance computing environment, a datacenter, a communications service provider infrastructure (e.g., one or more portions of an Evolved Packet Core), an in-memory computing environment, a computing system of a vehicle (e.g., an automobile or airplane), an Internet of Things (IoT) environment, an industrial control system, other computing environment, or combination thereof. In implementations herein, containers 320 may be implemented using hardware circuitry, such as one or more of a CPU, a GPU, a hardware accelerator, and so on. In one embodiment, containers 320 may be implemented using platform 202 described with respect to
Microservices containers 320 may include logic to implement the functionality of the microservice 325A, 325B (collectively referred to herein as microservices 325) and a sidecar 330A, 330B (collectively referred to herein as sidecars 330. A sidecar 330 can be a container that runs on the same pod as the microservice 325. As depicted herein, sidecar 330 is illustrated as part of the microservice container 320, but sidecar 330 may be implemented as a separate container then microservice 325 functionality in some implementations.
A local facilitator 310 is connected to the sidecars 330 and can operate in a privileged space of the microservice containers 320. In one implementation, local facilitator 310 is a privileged daemon with access to low-level information. For example, local facilitator 310 has access to low-level software telemetry and hardware data, such as registries.
In implementations herein, sidecar 330 may include one or more components to support synthetic data generation for enhanced microservice debugging in microservices architectures. These components can include telemetry data ingestion components 332A, 332B (collectively referred to herein as telemetry data ingestion component 332), collected data 334A, 334B (data stores collectively referred to as collected data 334), and microservice anomaly detector 336A, 336B (collectively referred to as microservice anomaly detector 336).
Service platform 300 also includes a service management component 340. Service management component 340 and its underlying sub-components may be implemented using hardware circuitry, such as one or more of a CPU, a GPU, a hardware accelerator, and so on. In one embodiment, service management component 340 may be implemented using platform 202 described with respect to
In one implementation, service management component 340 operates to control management and/or orchestration of resources, such as microservices, for a service of a service mesh hosted by a datacenter, such as datacenter system 100 of
Service management component 340 may include one or more components to support synthetic data generation for enhanced microservice debugging in microservices architectures. These components can include a controller 350, an evaluator 360, and a archive datastore 370. In implementations herein, the controller 350 can host a discovery and recommender service 352, an analytics manager 354, a synthetic data generator 356, a dynamic query generator 358, and an XPU manager 359. The evaluator can host an evaluation manager 362.
In implementations herein, the controller 350, evaluator 360, and archive datastore 370 may operate as part of a trusted execution environment (TEE) (not shown) generated by an underlying computing system(s) hosting the controller 350, evaluator 360, and archive datastore 370. The TEE may be authenticated and protected from unauthorized access using hardware support of the underlying computing system(s). Illustratively, the TEE may be embodied as one or more secure enclaves established using Intel™ SGX technology. The TEE may also include or otherwise interface with one or more drivers, libraries, or other components of the underlying computing system(s) to interface with one or more other XPUs.
In implementations herein, the microservice containers 320 and service management component 340 provide for synthetic data generation for enhanced microservice debugging in microservices architectures. In one implementation, the sidecar 330 for each microservice container 320 includes a telemetry data ingestion component 332 that receives telemetry data of the service platform 300 that is pertinent to the microservice 325. This telemetry data can include lower-level layers in the architecture (e.g., privileged space) and application (microservice 325) telemetry data and logs (e.g., user space). The collected data 334 is a data store that maintains this microservice-related telemetry data for the microservice 325.
The microservice anomaly detection component 336 analyzes telemetry data of the platform that is pertinent to the microservice 325. As noted above, this telemetry data can include lower-level layers in the service platform 300 (e.g., privileged space) and can include application (microservice 325) telemetry data and logs (e.g., user space). The microservice anomaly detection component 336 can provide hooks to capture errors in the applications (e.g., application service level objective (SLO) dictates processing at 30fps, but instead the application is processing at 28fps). In implementations herein, the microservice anomaly detection component 336 can consider infrastructure and application SLOs.
To obtain the collected data 334, the microservice anomaly detection component 336 can query the information available in the user space and, in addition, can invoke the local facilitator 310. The local facilitator 310 is connected to the sidecars 330 and has access to low level software telemetry and hardware data such as registries. As such, the local facilitator 310 can query the state of the service platform 300. Based on the collected data 334, the microservice anomaly detection component 336 can determine whether there are any deviations from normal or typical behavior. If a deviation is detection, the microservice anomaly detection component 336 can indicate the anomaly, including its type, to the service management component 340.
At the service management component 340, the controller 350 can provide for synthetic data generation that is utilized for dynamic query recommendations and ingestion for a given intricate interdependent set of microservices of the service hosted by the service platform 300. This synthetic data generation can be utilized with dynamic query recommendations and ingestion, which activates telemetry that is used to detect anomalies for enhanced debugging. The controller 350 can also discover an activation profile of the microservices 325 in order to provide recommendations to a service with respect to the information determined from the enhanced debugging provided by the synthetic data generation and dynamic query generation/injection in service platform 300.
In one implementation, controller 350 includes a discovery and recommender service 352 to discover and create an interdependency flow matrix for the service. The interdependency flow matrix may also be referred to as an interaction matrix or interactive matrix. The interdependency flow matrix can include a flow graph of services involving the interdependency of hardware devices (e.g., XPUs, IP blocks) and software services used in the service. In one embodiment, this interdependency flow matrix provides a virtual interaction matrix of all microservices of the service that are stitched together to provide the overall service offering of the application. In one implementation, the interdependency flow matrix is stored in archive datastore 370 of service management component 340.
The discovery and recommender service 352 may further generate recommendations on synthetic data generation and dynamic query generation for the service. In one implementation, the discovery and recommender service 352 may utilize the interdependency flow matrix in order to generate the recommendations on query generation.
In some implementations, the discovery and recommender service 352 may also access one or more activation profile(s) that are generated based on the telemetry data produced by the microservices 325. The activation profile(s) can represent a functional response to the query, where the response is produced by microservices 325 of the service when responding to the query. For example, an activation profile may identify the interfaces and compute elements (referred to herein as activation points) that are utilized by the service (and its deployed microservices) to provide a response to a query. The activation profile(s) can also include the sequence of responses by the activation points when responding to a query.
Using the information and recommendations from discovery and recommender service 352, the combination of analytics manager 354 and synthetic data generator 356 can provide for the synthetic data generation of implementations herein. For example, based on discovered activation profiles of the service, configured polices for the service, and the discovery and recommender service’s 352 inputs on a query to be exercised, the synthetic data generator 356 can generate a synthetic dataset based on machine learning (ML)-based heuristics of past real infield data provided by the service. The ML-based heuristics and generated filters can be applied to an original dataset of the service to generate privacy-filtered and aggregated data of the synthetic dataset. The analytics manager 354 can provide the ML-based heuristics and the filters to the synthetic data generator 356, as described in further detail below.
This analytics manager 354 performs ML analytics based on user-configured policy rules of the service. The analytics manager 354 can aggregates raw sensory data from a variety of input sources such as text, audio, video, image, gesture, motion, etc. This raw sensory input data is referred to herein a raw infield data, past infield data, original infield data, and so on. The analytics manager 354 may be communicatively coupled with the variety of input sources via an Internet Protocol network to aggregate data. Data aggregation policies (e.g., a sampling interval) used by the analytics manager 354 may be configurable in implementations herein.
The analytics manager 354 may host multiple sub-components including, but not limited to, a classification engine, an inference engine, a multiplex (mux)-demultiplex (demux) engine, a context engine, a semantics engine, and/or a privacy engine, to name a few examples.
The synthetic data generator 356 obtains trained models and/or a dataset from the analytics manager 354 and obtains user-configured profiles/templates for the service from storage, such as archive datastore 370. In some implementations, synthetic data generator 356 can identify when synthetic data may be useful for the service based on monitored activation profile(s) of the service and the interdependency flow matrix of the discovery and recommender service 352. The synthetic data generator 356 may utilize the user/administrator-defined profiles, policies, and templates of the service to generate the synthetic data.
In some implementations, the synthetic data generator 356 can load a synthetic data generator filter that is generated based on the obtained trained models and/or a dataset from the analytics manager 354 and/or based on the obtained user-configured profiles/templates for the service from storage. The synthetic data generator 356 may prioritize synthetic parameters in the filter based on service parameter used to model the microservices. In one implementation, the synthetic parameters can include, but are not limited to, privacy filters to apply to an original dataset of the service, an amount of noise to insert into a synthetic dataset, or a sampling interval. The service parameters can include quality of service (QoS) requirements and other parameters used to model the microservices. The synthetic data generator 356 may then generated the synthetic dataset for ingestion using the prioritized synthetic parameters in the filter. In one implementation, the synthetic dataset is generated by applying the synthetic parameters of the filter to an original infield dataset of the service. In some implementations, the synthetic dataset is generated based on revocation provisioning and configured policies for exclusions in the service.
In one implementation, the synthetic dataset generated from synthetic data generator 356 is provided to dynamic query generator 358. The dynamic query generator 358 can then create a new query for the service and/or modify an existing query to better target the microservices 325 of the service. In implementations herein, the generated dynamic query can be applied to the synthetic data generated by synthetic data generator 356. In some implementations, the query generated by dynamic query generator 358 can increase robustness of the service and microservices 325 by stress testing particular targeted component of the service.
In implementations herein, XPU manager 359 of controller 350 may provide for on-boarding and off-boarding of microservices 325 of the service based on the results of dynamic query injection as applied to the generated synthetic data discussed above. For example, the query injection and system reactions to the generated synthetic data can be captured as transactions that are logged into a blockchain and used at later points in time to diagnose any issues in the service. Based on this diagnosis, XPU manager 359 can cause microservices to be deployed or revoked to improve the robustness of the service in service platform 300. Furthermore, XPU manager 359 can utilize analysis of the transactions to inform future microservice deployments and service configurations for optimized operation of the service in the service platform 300.
In implementations herein, the evaluator 360 includes an evaluation manager 362 to provide a checks and balance to the controller 350. For example, the evaluation manager 362 can generate evaluation metrics based on whether the query recommendation and generation as applied to the generated synthetic data are generating metrics that are meeting quality and service level standards. In one example, the evaluation metrics are generated based on whether the synthetic dataset generates a synthetic activation profile that matches the one or more monitored previous activation profiles of the service. In one implementation, a synthetic activation profile that does not match the one or more monitored previous activation profiles of the service can be utilized for machine learning training for future anomaly detection and/or can be utilized to improve an algorithm of the machine learning training.
The evaluation metrics generated by evaluation manager 362 can act as a reward function that encourages positive behaviors from the controller 350 and discourages negative behaviors from the controller 350 (e.g., XPU manager 359 deploying microservices in line with query injection and results that meet quality and service standards). In some implementations, machine learning (ML)-based techniques can be applied for reward-based future improvements. As such, implementations herein provide a check and balances approach between the controller 350 and the evaluator 360.
Service management environment 400 includes an analytics manager 420, datastore 430, synthetic data generator 440, and dynamic query generator 450. In one implementation, analytics manager 420 is the same as analytics manager 354 of
In implementations herein, analytics manager 420 can perform ML analytics using user-configured policy rules 432 for a service. In one implementation, analytics manager 420 can aggregate raw sensory data (illustrated as past raw infield data 410) received from a variety of input sources. The variety of input source may include modalities such as text, audio, video, image, gesture, motion, and so on. In one implementation, the analytics manager 420 may be communicatively coupled with the variety of input sources via an IP network in order to receive and aggregate the past raw infield data 410. In implementations herein, the data aggregation policies (e.g., sampling intervals) applied by analytics manager 420 are configurable.
In implementations herein, the analytics manager 420 can include a classification engine 421, an inference engine 422, a mux/demux engine 423, a context engine 424, a semantics engine 425, and/or a privacy engine 426. More or less components than those illustrated and described with respect to
In one implementation, the classification engine 421 can perform rule-based data classification on the past raw infield data 410. For example, the past raw infield data 410 from the variety of input sources is classified with appropriate trained dataset for accurate inference. For example, audio data classification criteria could be orthogonal to image classification criteria.
The inference engine 422 can identify the characteristics in the aggregated data and performs rule based inference by co-relating the data across various sources. In one implementation, the inference engine 422 identifies such scenarios using the raw data, rules, user opt-in preferences (e.g., multi-sensor assertion based), and so on. In case of any conflict of sensor assertion, the inference engine 422 may use rule-based logic to resolve the conflicts. As an example, a lowest common denominator could be taken or an assertion score can be assigned to sensors based on the confidence of assertions in the past.
The mux/demux engine 423 can, based on configured privacy filters, perform protocol translation and framework-specific processing in order to get to the raw sensor data stream. Some data streams may be multiplexed, hence mux/demux engine 423 can be used for the same.
The context engine 424 can perform contextual analysis of the inferred raw sensory data. This may include association of context data across array of input sensory data (e.g., audio, activity, etc.) to determine an appropriate context of the input sensory data (e.g., running, talking, etc.).
The semantics engine 425 can perform semantics mapping to disambiguate events based on inferred context of the user and the sensory data. This aids the synthetic data generator 440 in applying appropriate templates that a user would like to communicate to the service platform (e.g., to the CSP).
The privacy engine 426 provides a privacy dashboard user interface (UI) for users to configure and manage privacy filters across the variety of sensors. The dashboard UI produces privacy ‘filters’ that are interposed the into sensor data stream in order to detect privacy-relevant content.
Utilizing the engines 421-416, the analytics manager 420 processing past raw infield data 410 to generate trained models and/or privacy filters 428. The synthetic data generator 440 can obtain the trained models/privacy filters 428 from analytics manager 420, as well as profiles and/or synthetic templates 434 for the service from the datastore 430. The synthetic data generator 440 can identify situations where synthetic data should be generated based on activation profile(s) and the interdependency flow matrix of the service. The synthetic data generator 440 can also use the user/admin-defined profiles, policies and templates 434 to generate the synthetic data.
In one example embodiment, using the trained models and/or privacy filters 428 generated by analytics manager 420, the synthetic data generator 440 can apply collaborative filtering and/or clustering to cluster a theme of a profile of the service. For example:
The dynamic query generator 450 may then generate a dynamic query, for example, as follows:
The above example dynamic query generation using synthetic data generation allows for debugging at scale in implementations herein.
Embodiments may be provided, for example, as a computer program product which may include one or more machine - readable media having stored thereon machine executable instructions that, when executed by one or more machines such as a computer, network of computers, or other electronic devices, may result in the one or more machines carrying out operations in accordance with embodiments described herein. A machine - readable medium may include, but is not limited to, floppy diskettes, optical disks, CD - ROMs (Compact Disc - Read Only Memories), and magneto - optical disks, ROMs, RAMS, EPROMs (Erasable Programmable Read Only Memories), EEPROMs (Electrically Erasable Programmable Read Only Memories), magnetic or optical cards, flash memory, or other type of media / machine - readable medium suitable for storing machine - executable instructions.
Moreover, embodiments may be downloaded as a computer program product, wherein the program may be transferred from a remote computer (e.g., a server) to a requesting computer (e.g., a client) by way of one or more data signals embodied in and / or modulated by a carrier wave or other propagation medium via a communication link (e.g., a modem and / or network connection).
Throughout the document, term “user” may be interchangeably referred to as “viewer”, “observer”, “speaker”, “person”, “individual”, “end - user”, and / or the like. It is to be noted that throughout this document, terms like “graphics domain” may be referenced interchangeably with “graphics processing unit”, “graphics processor”, or simply “GPU” and similarly, “CPU domain” or “host domain” may be referenced interchangeably with “computer processing unit”, “application processor”, or simply “CPU” .
It is to be noted that terms like “node”, “computing node”, “server”, “server device”, “cloud computer”, “cloud server”, “cloud server computer”, “machine”, “host machine”, “device”, “computing device”, “computer”, “computing system”, and the like, may be used interchangeably throughout this document. It is to be further noted that terms like “application”, “software application”, “program”, “software program”, “package”, “software package”, and the like, may be used interchangeably throughout this document. Also, terms like “job”, “input”, “request”, “message ”, and the like, may be used interchangeably throughout this document.
The process of method 500 is illustrated in linear sequences for brevity and clarity in presentation; however, it is contemplated that any number of them can be performed in parallel, asynchronously, or in different orders. Further, for brevity, clarity, and ease of understanding, many of the components and processes described with respect to
The example process of method 500 of
Subsequently, at block 530, the processing device may prioritize synthetic parameters based on service parameter used to model the microservices. In one implementation, the service parameters can include QoS requirements and other parameters used to model the microservices. Then, at block 540, the processing device may generate a synthetic dataset for ingestion using prioritized synthetic parameters in the filter. In one implementation, the synthetic dataset is generated by applying the synthetic parameters of the filter to an original infield dataset of the service. In some implementations, the synthetic dataset is generated based on revocation provisioning and configured policies for exclusions in the service.
At block 550, the processing device may, in response to receiving a synthetic activation profile generated using the synthetic dataset that matches one or more of the monitored previous activation profiles, demux the generated synthetic dataset corresponding to the synthetic activation profile and perform reverse mapping on the demuxed synthetic dataset to match to original data in the original infield dataset. Lastly, at block 555, the processing device may cause a policy-based action to be performed with respect to the original data mapped to the original infield dataset and continue execution with a current configuration.
The process of method 560 is illustrated in linear sequences for brevity and clarity in presentation; however, it is contemplated that any number of them can be performed in parallel, asynchronously, or in different orders. Further, for brevity, clarity, and ease of understanding, many of the components and processes described with respect to
The example process of method 560 of
Subsequently, at block 575, the processing device may verify that a dataset is available for new configured polices. Lastly, at block 580, the processing device may enforce configured policy with an appropriate mux/demux channel configuration.
The computing device 600 may additionally include one or more of the following: cache 662, a graphical processing unit (GPU) 612 (which may be the hardware accelerator in some implementations), a wireless input/output (I/O) interface 620, a wired I/O interface 630, system memory 640 (e.g., memory circuitry), power management circuitry 650, non-transitory storage device 660, and a network interface 670 for connection to a network 672. The following discussion provides a brief, general description of the components forming the illustrative computing device 600. Example, non-limiting computing devices 600 may include a desktop computing device, blade server device, workstation, or similar device or system.
In embodiments, the processor cores 618 are capable of executing machine-readable instruction sets 614, reading data and/or instruction sets 614 from one or more storage devices 660 and writing data to the one or more storage devices 660. Those skilled in the relevant art can appreciate that the illustrated embodiments as well as other embodiments may be practiced with other processor-based device configurations, including portable electronic or handheld electronic devices, for instance smartphones, portable computers, wearable computers, consumer electronics, personal computers (“PCs”), network PCs, minicomputers, server blades, mainframe computers, and the like.
The processor cores 618 may include any number of hardwired or configurable circuits, some or all of which may include programmable and/or configurable combinations of electronic components, semiconductor devices, and/or logic elements that are disposed partially or wholly in a PC, server, or other computing system capable of executing processor-readable instructions.
The computing device 600 includes a bus or similar communications link 616 that communicably couples and facilitates the exchange of information and/or data between various system components including the processor cores 618, the cache 662, the graphics processor circuitry 612, one or more wireless I/O interfaces 620, one or more wired I/O interfaces 630, one or more storage devices 660, and/or one or more network interfaces 670. The computing device 600 may be referred to in the singular herein, but this is not intended to limit the embodiments to a single computing device 600, since in certain embodiments, there may be more than one computing device 600 that incorporates, includes, or contains any number of communicably coupled, collocated, or remote networked circuits or devices.
The processor cores 618 may include any number, type, or combination of currently available or future developed devices capable of executing machine-readable instruction sets.
The processor cores 618 may include (or be coupled to) but are not limited to any current or future developed single- or multi-core processor or microprocessor, such as: on or more systems on a chip (SOCs); central processing units (CPUs); digital signal processors (DSPs); graphics processing units (GPUs); application-specific integrated circuits (ASICs), programmable logic units, field programmable gate arrays (FPGAs), and the like. Unless described otherwise, the construction and operation of the various blocks shown in
The system memory 640 may include read-only memory (“ROM”) 642 and random access memory (“RAM”) 646. A portion of the ROM 642 may be used to store or otherwise retain a basic input/output system (“BIOS”) 644. The BIOS 644 provides basic functionality to the computing device 600, for example by causing the processor cores 618 to load and/or execute one or more machine-readable instruction sets 614. In embodiments, at least some of the one or more machine-readable instruction sets 614 cause at least a portion of the processor cores 618 to provide, create, produce, transition, and/or function as a dedicated, specific, and particular machine, for example a word processing machine, a digital image acquisition machine, a media playing machine, a gaming system, a communications device, a smartphone, or similar.
The computing device 600 may include at least one wireless input/output (I/O) interface 620. The at least one wireless I/O interface 620 may be communicably coupled to one or more physical output devices 622 (tactile devices, video displays, audio output devices, hardcopy output devices, etc.). The at least one wireless I/O interface 620 may communicably couple to one or more physical input devices 624 (pointing devices, touchscreens, keyboards, tactile devices, etc.). The at least one wireless I/O interface 620 may include any currently available or future developed wireless I/O interface. Example wireless I/O interfaces include, but are not limited to: BLUETOOTH®, near field communication (NFC), and similar.
The computing device 600 may include one or more wired input/output (I/O) interfaces 630. The at least one wired I/O interface 630 may be communicably coupled to one or more physical output devices 622 (tactile devices, video displays, audio output devices, hardcopy output devices, etc.). The at least one wired I/O interface 630 may be communicably coupled to one or more physical input devices 624 (pointing devices, touchscreens, keyboards, tactile devices, etc.). The wired I/O interface 630 may include any currently available or future developed I/O interface. Example wired I/O interfaces include, but are not limited to: universal serial bus (USB), IEEE 1394 (“FireWire”), and similar.
The computing device 600 may include one or more communicably coupled, non-transitory, data storage devices 660. The data storage devices 660 may include one or more hard disk drives (HDDs) and/or one or more solid-state storage devices (SSDs). The one or more data storage devices 660 may include any current or future developed storage appliances, network storage devices, and/or systems. Nonlimiting examples of such data storage devices 660 may include, but are not limited to, any current or future developed non-transitory storage appliances or devices, such as one or more magnetic storage devices, one or more optical storage devices, one or more electro-resistive storage devices, one or more molecular storage devices, one or more quantum storage devices, or various combinations thereof. In some implementations, the one or more data storage devices 660 may include one or more removable storage devices, such as one or more flash drives, flash memories, flash storage units, or similar appliances or devices capable of communicable coupling to and decoupling from the computing device 600.
The one or more data storage devices 660 may include interfaces or controllers (not shown) communicatively coupling the respective storage device or system to the bus 616. The one or more data storage devices 660 may store, retain, or otherwise contain machine-readable instruction sets, data structures, program modules, data stores, databases, logical structures, and/or other data useful to the processor cores 618 and/or graphics processor circuitry 612 and/or one or more applications executed on or by the processor cores 618 and/or graphics processor circuitry 612. In some instances, one or more data storage devices 660 may be communicably coupled to the processor cores 618, for example via the bus 616 or via one or more wired communications interfaces 630 (e.g., Universal Serial Bus or USB); one or more wireless communications interfaces 620 (e.g., Bluetooth®, Near Field Communication or NFC); and/or one or more network interfaces 670 (IEEE 802.3 or Ethernet, IEEE 802.11, or Wi-Fi®, etc.).
Processor-readable instruction sets 614 and other programs, applications, logic sets, and/or modules may be stored in whole or in part in the system memory 640. Such instruction sets 614 may be transferred, in whole or in part, from the one or more data storage devices 660. The instruction sets 614 may be loaded, stored, or otherwise retained in system memory 640, in whole or in part, during execution by the processor cores 618 and/or graphics processor circuitry 612.
The computing device 600 may include power management circuitry 650 that controls one or more operational aspects of the energy storage device 652. In embodiments, the energy storage device 652 may include one or more primary (i.e., non-rechargeable) or secondary (i.e., rechargeable) batteries or similar energy storage devices. In embodiments, the energy storage device 652 may include one or more supercapacitors or ultracapacitors. In embodiments, the power management circuitry 650 may alter, adjust, or control the flow of energy from an external power source 654 to the energy storage device 652 and/or to the computing device 600. The power source 654 may include, but is not limited to, a solar power system, a commercial electric grid, a portable generator, an external energy storage device, or any combination thereof.
For convenience, the processor cores 618, the graphics processor circuitry 612, the wireless I/O interface 620, the wired I/O interface 630, the storage device 660, and the network interface 670 are illustrated as communicatively coupled to each other via the bus 616, thereby providing connectivity between the above-described components. In alternative embodiments, the above-described components may be communicatively coupled in a different manner than illustrated in
The following examples pertain to further embodiments. Example 1 is an apparatus to facilitate synthetic data generation for enhanced microservice debugging. The apparatus of Example 1 comprises one or more processors to: load a filter for a synthetic data generator for a service deployed in a datacenter system, the filter configured for the service based on service policies; prioritize synthetic parameters of the filter based on service parameters used to model microservices deployed for the service; generate a synthetic dataset for ingestion using the prioritized synthetic parameters in the filter of the synthetic data generator, the synthetic dataset generated by applying the synthetic parameters of the filter to an original infield dataset of the service; demultiplex the synthetic dataset in response to a synthetic activation profile generated using the synthetic dataset matching one or more monitored previous activation profiles of the service; and reverse map the demultiplexed synthetic dataset to match to original data in the original infield dataset.
In Example 2, the subject matter of Example 1 can optionally include wherein the one or more processors are further to cause a policy-based action to occur with respect to the original data mapped to the original infield dataset. In Example 3, the subject matter of any one of Examples 1-2 can optionally include wherein the service parameters comprise quality of service (QoS) metrics configured for the service. In Example 4, the subject matter of any one of Examples 1-3 can optionally include wherein the synthetic parameters comprise privacy filters to apply to an original dataset of the service, an amount of noise to insert into a synthetic dataset, or a sampling interval.
In Example 5, the subject matter of any one of Examples 1-4 can optionally include wherein the one or more processors provide a trusted execution environment (TEE) for a controller of the service to generate the synthetic dataset using the synthetic data generator. In Example 6, the subject matter of any one of Examples 1-5 can optionally include wherein the filter is configured for the service by the one or more processors to: provide a user interface (UI) to a user, the UI comprising options for telemetry activation parameters based on the service policies; receiving, via the UI, configurations for the options for the telemetry activation parameters; and generating the filter based on the configurations that are received for the options.
In Example 7, the subject matter of any one of Examples 1-6 can optionally include wherein generating the filter further comprises the one or more processors to apply a machine learning analytics engine to the original infield dataset to generate a trained model based on data aggregation using one or more of classification, inference, contextual analysis, or semantics mapping, the trained model used as the filter to generate the synthetic dataset at the synthetic data generator. In Example 8, the subject matter of any one of Examples 1-7 can optionally include wherein the synthetic dataset is ingested by a synthetic query generator of the service, the synthetic query generator to generated synthetic queries for the service using the synthetic dataset.
In Example 9, the subject matter of any one of Examples 1-8 can optionally include wherein the one or processors are to generate the synthetic dataset based on revocation provisioning and configured policies for exclusions in the service. In Example 10, the subject matter of any one of Examples 1-9 can optionally include wherein the one or more processors are further to generate one or more evaluation metrics based on whether the synthetic dataset generates a synthetic activation profile that matches the one or more monitored previous activation profiles of the service, wherein the synthetic activation profile that does not match the one or more monitored previous activation profiles of the service is utilized for machine learning training for future anomaly detection or to improve an x of the machine learning training.
Example 11 is a non-transitory computer-readable storage medium for facilitating synthetic data generation for enhanced microservice debugging. The non-transitory computer-readable storage medium of Example 11 having stored thereon executable computer program instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising: loading, by the one or more processors, a filter for a synthetic data generator for a service deployed in a datacenter system, the filter configured for the service based on service policies; prioritizing, by the one or more processors, synthetic parameters of the filter based on service parameters used to model microservices deployed for the service; generating, by the one or more processors, a synthetic dataset for ingestion using the prioritized synthetic parameters in the filter of the synthetic data generator, the synthetic dataset generated by applying the synthetic parameters of the filter to an original infield dataset of the service; demultiplexing, by the one or more processors, the synthetic dataset in response to a synthetic activation profile generated using the synthetic dataset matching one or more monitored previous activation profiles of the service; and reverse mapping, by the one or more processors, the demultiplexed synthetic dataset to match to original data in the original infield dataset.
In Example 12, the subject matter of Example 11 can optionally include wherein the operations further comprise causing a policy-based action to occur with respect to the original data mapped to the original infield dataset. In Example 13, the subject matter of Examples 11-12 can optionally include wherein the synthetic parameters comprise privacy filters to apply to an original dataset of the service, an amount of noise to insert into a synthetic dataset, or a sampling interval.
In Example 14, the subject matter of Examples 11-13 can optionally include wherein the filter is configured for the service by the one or more processors to: provide a user interface (UI) to a user, the UI comprising options for telemetry activation parameters based on the service policies; receiving, via the UI, configurations for the options for the telemetry activation parameters; and generating the filter based on the configurations that are received for the options. In Example 15, the subject matter of Examples 11-14 can optionally include wherein generating the filter further comprises the one or more processors to apply a machine learning analytics engine to the original infield dataset to generate a trained model based on data aggregation using one or more of classification, inference, contextual analysis, or semantics mapping, the trained model used as the filter to generate the synthetic dataset at the synthetic data generator.
Example 16 is a method for facilitating synthetic data generation for enhanced microservice debugging. The method of Example 16 can include loading, by one or more processors, a filter for a synthetic data generator for a service deployed in a datacenter system, the filter configured for the service based on service policies; prioritizing, by the one or more processors, synthetic parameters of the filter based on service parameters used to model microservices deployed for the service; generating, by the one or more processors, a synthetic dataset for ingestion using the prioritized synthetic parameters in the filter of the synthetic data generator, the synthetic dataset generated by applying the synthetic parameters of the filter to an original infield dataset of the service; demultiplexing, by the one or more processors, the synthetic dataset in response to a synthetic activation profile generated using the synthetic dataset matching one or more monitored previous activation profiles of the service; and reverse mapping, by the one or more processors, the demultiplexed synthetic dataset to match to original data in the original infield dataset.
In Example 17, the subject matter of Example 16 can optionally include further comprising causing a policy-based action to occur with respect to the original data mapped to the original infield dataset. In Example 18, the subject matter of Examples 16-17 can optionally include wherein the synthetic parameters comprise privacy filters to apply to an original dataset of the service, an amount of noise to insert into a synthetic dataset, or a sampling interval.
In Example 19, the subject matter of Examples 16-18 can optionally include wherein the filter is configured for the service by the one or more processors to: provide a user interface (UI) to a user, the UI comprising options for telemetry activation parameters based on the service policies; receiving, via the UI, configurations for the options for the telemetry activation parameters; and generating the filter based on the configurations that are received for the options. In Example 20, the subject matter of Examples 16-19 can optionally include wherein generating the filter further comprises the one or more processors to apply a machine learning analytics engine to the original infield dataset to generate a trained model based on data aggregation using one or more of classification, inference, contextual analysis, or semantics mapping, the trained model used as the filter to generate the synthetic dataset at the synthetic data generator.
Example 21 is a system for facilitating synthetic data generation for enhanced microservice debugging. The system of Example 21 can optionally include a memory to store a block of data, and a processor communicably coupled to the memory to: load a filter for a synthetic data generator for a service deployed in a datacenter system, the filter configured for the service based on service policies; prioritize synthetic parameters of the filter based on service parameters used to model microservices deployed for the service; generate a synthetic dataset for ingestion using the prioritized synthetic parameters in the filter of the synthetic data generator, the synthetic dataset generated by applying the synthetic parameters of the filter to an original infield dataset of the service; demultiplex the synthetic dataset in response to a synthetic activation profile generated using the synthetic dataset matching one or more monitored previous activation profiles of the service; and reverse map the demultiplexed synthetic dataset to match to original data in the original infield dataset.
In Example 22, the subject matter of Example 21 can optionally include wherein the one or more processors are further to cause a policy-based action to occur with respect to the original data mapped to the original infield dataset. In Example 23, the subject matter of any one of Examples 21-22 can optionally include wherein the service parameters comprise quality of service (QoS) metrics configured for the service. In Example 24, the subject matter of any one of Examples 21-23 can optionally include wherein the synthetic parameters comprise privacy filters to apply to an original dataset of the service, an amount of noise to insert into a synthetic dataset, or a sampling interval.
In Example 25, the subject matter of any one of Examples 21-24 can optionally include wherein the one or more processors provide a trusted execution environment (TEE) for a controller of the service to generate the synthetic dataset using the synthetic data generator. In Example 26, the subject matter of any one of Examples 21-25 can optionally include wherein the filter is configured for the service by the one or more processors to: provide a user interface (UI) to a user, the UI comprising options for telemetry activation parameters based on the service policies; receiving, via the UI, configurations for the options for the telemetry activation parameters; and generating the filter based on the configurations that are received for the options.
In Example 27, the subject matter of any one of Examples 21-26 can optionally include wherein generating the filter further comprises the one or more processors to apply a machine learning analytics engine to the original infield dataset to generate a trained model based on data aggregation using one or more of classification, inference, contextual analysis, or semantics mapping, the trained model used as the filter to generate the synthetic dataset at the synthetic data generator. In Example 28, the subject matter of any one of Examples 21-27 can optionally include wherein the synthetic dataset is ingested by a synthetic query generator of the service, the synthetic query generator to generated synthetic queries for the service using the synthetic dataset.
In Example 29, the subject matter of any one of Examples 21-28 can optionally include wherein the one or processors are to generate the synthetic dataset based on revocation provisioning and configured policies for exclusions in the service. In Example 30, the subject matter of any one of Examples 21-29 can optionally include wherein the one or more processors are further to generate one or more evaluation metrics based on whether the synthetic dataset generates a synthetic activation profile that matches the one or more monitored previous activation profiles of the service, wherein the synthetic activation profile that does not match the one or more monitored previous activation profiles of the service is utilized for machine learning training for future anomaly detection or to improve an x of the machine learning training.
Example 30 is an apparatus for facilitating synthetic data generation for enhanced microservice debugging, comprising means for loading a filter for a synthetic data generator for a service deployed in a datacenter system, the filter configured for the service based on service policies; means for prioritizing synthetic parameters of the filter based on service parameters used to model microservices deployed for the service; means for generating a synthetic dataset for ingestion using the prioritized synthetic parameters in the filter of the synthetic data generator, the synthetic dataset generated by applying the synthetic parameters of the filter to an original infield dataset of the service; means for demultiplexing the synthetic dataset in response to a synthetic activation profile generated using the synthetic dataset matching one or more monitored previous activation profiles of the service; and means for reverse mapping the demultiplexed synthetic dataset to match to original data in the original infield dataset. In Example 31, the subject matter of Example 30 can optionally include the apparatus further configured to perform the method of any one of the Examples 17 to 20.
Example 32 is at least one machine readable medium comprising a plurality of instructions that in response to being executed on a computing device, cause the computing device to carry out a method according to any one of Examples 16-20. Example 33 is an apparatus for facilitating synthetic data generation for enhanced microservice debugging, configured to perform the method of any one of Examples 16-20. Example 34 is an apparatus for facilitating synthetic data generation for enhanced microservice debugging, comprising means for performing the method of any one of claims 16 to 20. Specifics in the Examples may be used anywhere in one or more embodiments.
The foregoing description and drawings are to be regarded in an illustrative rather than a restrictive sense. Persons skilled in the art can understand that various modifications and changes may be made to the embodiments described herein without departing from the broader spirit and scope of the features set forth in the appended claims.
