Patent Application
20230125491
The technical character of the present disclosure generally relates to the field of Information Technology (IT) systems, and more particularly, systems and methods for workload migration in IT systems.
In the field of IT, workload migration describes moving one or more IT systems with all associated infrastructure and applications from one operating environment to another. Typically, a workload migration program is driven by an external compelling event such as data center closure. The execution of a workload migration program is usually a complex problem due to many factors, including: the large number of components and the relationships between all the Configuration Items (CIs); the time required for workload migration often means that business changes must be accommodated (which may impact CI relationships and dependencies); the involvement of multiple data centers; and the disruptive nature of workload migration.
The typical solution to discovery of the CIs and their relationships is to place agents on endpoint devices or run agentless scans from centralized systems and gather information about installed components and inspect network flows to determine relationships. This raw data is then sent back to centralized systems for analysis and to support workload migration planning activities. The scans are repeated to ensure changes to the IT environment are captured. Such a known solution is resource intensive and time-consuming, and thus improved approaches to workload disruption are desired.
Aspects of the present invention provide a method of training a machine learning algorithm for identifying an edge computing environment location as a target for workload migration. Aspects of the present invention also provide a method for workload migration at an edge computing environment. Such methods may be computer-implemented. That is, such methods may be implemented in a computer infrastructure having computer executable code tangibly embodied on a computer readable storage medium having programming instructions configured to perform a proposed method. Aspects of the present invention further provide a computer program product including computer program code for implementing the proposed concepts when executed on a processor. Aspects of the present invention yet further provide a system for training a machine learning algorithm for identifying an edge computing environment location as a target for workload migration. Aspects of the present invention also provide a system for workload migration at an edge computing environment.
According to an aspect of the present invention, there is a method of training a machine learning algorithm for identifying an edge computing environment location as a target for workload migration. The method comprises receiving: i) a description of a training workload; and ii) an identifier of an edge location for migrating the training workload. The method also comprises training the machine learning algorithm based on the description of a training workload as a training input for the machine learning algorithm and the identifier of the edge location for migrating the training workload as a training output for the machine learning algorithm.
Implementations of the invention may utilize edge computing. Such implementations may leverage reduced processing and power costs of edge locations which can be positioned: close to the systems to be migrated; on the network(s); and/or embedded with the workloads to be migrated. Embodiments may thus make us of edge locations that are better positioned to detect and react to changes. Also, discovery and analytics may be distributed across multiple edge locations.
Embodiments may thus provide concepts for workload migration execution that may facilitate: reduced costs; higher quality; and/or reduced disruption.
In particular, embodiments may propose to use edge locations as sources for migrating workload between data centers.
In embodiments, to identify the latest (i.e., most up-to-date version) of workloads, suitable times for migration, and/or an aggregated group of dependent workloads, implementations of the invention use machine learning algorithms (e.g., deep learning models) that may be deployed in respective edge locations. Usage of distributed learning may make embodiments efficient, scalable and/or optimal. For instance, learning from different edge locations may further contribute to a more efficient centralized knowledge storage (where learning can be stored in the form of metadata for example), based on summary and analysis of raw data captured at numerous edge locations. Such learning and knowledge may improve a proposed workload migration model.
By way of example, the training may comprise adjusting a configuration of the machine learning algorithm so as to decrease an outage time objective. The proposed training concept(s) may thus be tailored to specific requirements or targets.
In some embodiments, the training may comprise adjusting a configuration of the machine learning algorithm so as to decrease a difference between the identifier of the edge location for migrating the training workload and a second identifier that is output by the machine learning component. That is, the machine learning algorithm may be trained towards outputting a known output for a known input, thus making use of known/training knowledge to improve prediction/output accuracy. The machine learning algorithm may thus be trained using a corpus of training workloads and identifiers of edge locations for migrating the training workloads.
The machine learning algorithm may, for example, comprise a neural network wherein the training is based on a loss function for the gradient of the neural network.
In some embodiments, the description of the training workload comprises a time series of a plurality of variables.
According to another aspect of the present invention, there is a method for workload migration at an edge computing environment. The method comprises providing a description of a workload to a machine learning algorithm for identifying an edge computing environment location as a target for workload migration. The method also comprises obtaining a prediction result from the machine learning algorithm based on the description of the workload, the prediction result including an identifier of an edge location.
In some implementations, a machine learning algorithm is trained using a training algorithm configured to receive an array of training inputs and known outputs, wherein the training inputs comprise descriptions of training workloads and the known outputs comprise identifiers of edge locations for migrating the training workloads.
Some embodiments may be configured to communicate the workload to an edge computing environment associated with the identifier of an edge location included in the prediction result.
Embodiments may be employed in combination with conventional/existing workload management systems and/or workload migration systems. In this way, embodiments may integrate into legacy systems so as to improve and/or extend their functionality and capabilities. An improved workload management/migration system may therefore be provided by proposed embodiments.
According to another embodiment of the present invention, there is a computer program product for training a machine learning algorithm for identifying an edge computing environment location as a target for workload migration, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processing unit to cause the processing unit to perform a method according to one or more proposed embodiments when executed on at least one processor of a data processing system.
According to another embodiment of the present invention, there is a computer program product for workload migration at an edge computing environment, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processing unit to cause the processing unit to perform a method according to one or more proposed embodiments when executed on at least one processor of a data processing system.
According to yet another aspect, there is a processing system comprising at least one processor and the computer program product according to one or more embodiments, wherein the at least one processor is adapted to execute the computer program code of said computer program product.
According to another aspect, there is a system for training a machine learning algorithm for identifying an edge computing environment location as a target for workload migration. The system comprises a processor arrangement configured to perform the steps of: receiving: i) a description of a training workload; and ii) an identifier of an edge location for migrating the training workload; and training the machine learning algorithm based on the description of a training workload as a training input for the machine learning algorithm and the identifier of the edge location for migrating the training workload as a training output for the machine learning algorithm.
According to yet another aspect, there is a system for workload migration at an edge computing environment. The system comprises a processor arrangement configured to perform the steps of: providing a description of a workload to a machine learning algorithm for identifying an edge computing environment location as a target for workload migration; and obtaining a prediction result from the machine learning algorithm based on the description of the workload, the prediction result including an identifier of an edge location.
In this manner, implementations of the invention provide machine learning approaches for identifying one or more edge locations as the source(s) for each workload wave during migration, disaster recovery, or redundancy from one compute facility (whether temporary, mobile, closet, or datacenter) to another. For instance, embodiments may provide for a machine-learning algorithm to be trained to nominate or suggest the most appropriate edge location(s) for migration of a workload. Using a description of a workload, the machine-learning algorithm may identify/suggest one or more edge locations for migrating the workload.
Aspects of the present invention are described in the detailed description which follows, in reference to the noted plurality of drawings by way of non-limiting examples of exemplary embodiments of the present invention.
It should be understood that the Figures are merely schematic and are not drawn to scale. It should also be understood that the same reference numerals are used throughout the Figures to indicate the same or similar parts.
In the context of the present disclosure, where embodiments of the present invention constitute a method, it should be understood that such a method is a process for execution by a computer, i.e., is a computer-implementable method. The various steps of the method therefore reflect various parts of a computer program, e.g., various parts of one or more algorithms.
Also, in the context of the present disclosure, a (processing) system may be a single device or a collection of distributed devices that are adapted to execute one or more embodiments of the methods of the present invention. For instance, a system may be a personal computer (PC), a server or a collection of PCs and/or servers connected via a network such as a local area network, the Internet and so on to cooperatively execute at least one embodiment of the methods of the present invention.
Also, in the context of the present disclosure, a system may be a single device or a collection of distributed devices that are adapted to execute one or more embodiments of the methods of the present invention. For instance, a system may be a personal computer (PC), a portable computing device (such as a tablet computer, laptop, smartphone, etc.), a set-top box, a server or a collection of PCs, IoT sensors and/or servers connected via a network such as a local area network, 5G networking (i.e., mobile device networking), the Internet and so on to cooperatively execute at least one embodiment of the methods of the present invention.
The technical character of the present disclosure generally relates to workload management, and more particularly, to workload migration that may, for example, identify edge computing environments (i.e., edge locations) for migrating one or more workloads. More specifically, embodiments of the present invention provide concepts for automatically identifying an edge computing environment location as a target for workload migration, and such identification may employ machine learning techniques to analyze workload information (and suggest/predict one or more edge computing environments (i.e., edge locations) for migrating the workload.
Edge computing is a distributed computing paradigm which brings computation and data storage closer to the location it is needed, to improve response times and save bandwidth. In this paradigm, the smaller infrastructure components are distributed at the edge of the network. This facilitates the operation of end devices by acting as the relay to meet the needs of high-speed Internet of Things devices, thus reducing the bandwidth load of the network core. The same paradigm can be applied to cloud computing when it comes to connected devices. In such an arrangement, the first connection (i.e., hop) from smart devices, mobile phones, connected cars, drones, robots, etc. is not to large public clouds, but is instead to something closer, the Edge Cloud operating at the edge of the network. Unlike conventional workload migration approaches, embodiments may propose concepts for intelligently determining which edge computing resources (i.e., edge locations) to migrate one or more workloads to.
Embodiments may also be able to adapt to feedback and/or changing resources for future instances/occurrences of the same workload(s).
Embodiments may therefore obviate or mitigate problems associated with conventional workload migration approaches, by providing a method, a system and a computer program product for workload migration at an edge computing environment using machine learning algorithms. Also, concepts for training the machine-learning algorithms are provided.
Implementations may achieve such benefits through the use of machine-learning training approaches employing a corpus of training workloads and edge locations for migrating the training workloads. In some embodiments, for example, training may comprise adjusting a configuration of the machine learning algorithm so as to decrease a difference between the identifier of the edge location for migrating the training workload and a second identifier that is output by the machine learning component.
By way of further example, training may comprise adjusting a configuration of the machine learning algorithm so as to decrease an outage time objective.
After training, the machine learning algorithm may be expected to be able to receive a description of a workload (e.g., a time series of a plurality of variables) and provide a prediction/suggestion of an edge location most suited for migrating the workload.
It is understood in advance that although this disclosure includes a detailed description on cloud computing, implementation of the techniques recited herein are not limited to a cloud computing environment. Rather, embodiments of the present invention are capable of being implemented in conjunction with any other type of computing environment now known or later developed.
Cloud computing is a model of service delivery for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, network bandwidth, servers, processing, memory, storage, applications, virtual machines, and services) that can be rapidly provisioned and released with minimal management effort or interaction with a provider of the service. This cloud model may include at least five characteristics, at least three service models, and at least four deployment models.
Characteristics are as follows:
Service Models are as follows:
Deployment Models are as follows:
A cloud computing environment is service oriented with a focus on statelessness, low coupling, modularity, and semantic interoperability. At the heart of cloud computing is an infrastructure comprising a network of interconnected nodes.
Referring now to
In cloud computing node 10 there is a computer system/server 12, which is operational with numerous other general purpose or special purpose computing system environments or configurations. Examples of well-known computing systems, environments, and/or configurations that may be suitable for use with computer system/server 12 include, but are not limited to, personal computer systems, server computer systems, thin clients, thick clients, handheld or laptop devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable consumer electronics, network PCs, minicomputer systems, mainframe computer systems, and distributed cloud computing environments that include any of the above systems or devices, and the like.
Computer system/server 12 may be described in the general context of computer system-executable instructions, such as program modules, being executed by a computer system. Generally, program modules may include routines, programs, objects, components, logic, data structures, and so on that perform particular tasks or implement particular abstract data types. Computer system/server 12 may be practiced in distributed cloud computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed cloud computing environment, program modules may be located in both local and remote computer system storage media including memory storage devices.
As shown in
Bus 18 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. By way of example, and not limitation, such architectures include Industry Standard Architecture (ISA) bus, Micro Channel Architecture (MCA) bus, Enhanced ISA (EISA) bus, Video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnects (PCI) bus.
Computer system/server 12 typically includes a variety of computer system readable media. Such media may be any available media that is accessible by computer system/server 12, and it includes both volatile and non-volatile media, removable and non-removable media.
System memory 28 can include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and/or cache memory 32. Computer system/server 12 may further include other removable/non-removable, volatile/non-volatile computer system storage media. By way of example only, storage system 34 can be provided for reading from and writing to a non-removable, non-volatile magnetic media (not shown and typically called a “hard drive”). Although not shown, a magnetic disk drive for reading from and writing to a removable, non-volatile magnetic disk (e.g., a “floppy disk”), and an optical disk drive for reading from or writing to a removable, non-volatile optical disk such as a CD-ROM, DVD-ROM or other optical media can be provided. In such instances, each can be connected to bus 18 by one or more data media interfaces. As will be further depicted and described below, memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the invention.
Program/utility 40, having a set (at least one) of program modules 42, may be stored in memory 28 by way of example, and not limitation, as well as an operating system, one or more application programs, other program modules, and program data. Each of the operating system, one or more application programs, other program modules, and program data or some combination thereof, may include an implementation of a networking environment.
Program modules 42 generally carry out the functions and/or methodologies of embodiments of the invention as described herein. For example, some or all of the functions of a DHCP client 80 can be implemented as one or more of the program modules 42. Additionally, the DHCP client 80 may be implemented as separate dedicated processors or a single or several processors to provide the functionality described herein. In embodiments, the DHCP client 80 performs one or more of the processes described herein.
Computer system/server 12 may also communicate with one or more external devices 14 such as a keyboard, a pointing device, a display 24, etc.; one or more devices that enable a user to interact with computer system/server 12; and/or any devices (e.g., network card, modem, etc.) that enable computer system/server 12 to communicate with one or more other computing devices. Such communication can occur via I/O interfaces 22. Still yet, computer system/server 12 can communicate with one or more networks such as a local area network (LAN), a general wide area network (WAN), and/or a public network (e.g., the Internet) via network adapter 20. As depicted, network adapter 20 communicates with the other components of computer system/server 12 via bus 18. It should be understood that although not shown, other hardware and/or software components could be used in conjunction with computer system/server 12. Examples, include, but are not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID (redundant array of inexpensive disks or redundant array of independent disks) systems, tape drives, and data archival storage systems, etc.
Referring now to
Referring now to
Hardware and software layer 60 includes hardware and software components. Examples of hardware components include mainframes 61; RISC (Reduced Instruction Set Computer) architecture based servers 62; servers 63; blade servers 64; storage device 65; and networks and networking components 66. In some embodiments, software components include network application server software 67 and database software 68.
Virtualization layer 70 provides an abstraction layer from which the following examples of virtual entities may be provided: virtual servers 71; virtual storage 72; virtual networks 73, including virtual private networks; virtual applications and operating systems 74; and virtual clients 75.
In one example, management layer 80 may provide the functions described below. Resource provisioning 81 provides dynamic procurement of computing resources and other resources that are utilized to perform tasks within the cloud computing environment. Metering and Pricing 82 provide cost tracking as resources are utilized within the cloud computing environment, and billing or invoicing for consumption of these resources. In one example, these resources may comprise application software licenses. Security provides identity verification for cloud consumers and tasks, as well as protection for data and other resources. User portal 83 provides access to the cloud computing environment for consumers and system administrators. Service level management 84 provides cloud computing resource allocation and management such that required service levels are met. Service Level Agreement (SLA) planning and fulfillment 85 provide pre-arrangement for, and procurement of, cloud computing resources for which a future requirement is anticipated in accordance with an SLA.
Workloads layer 90 provides examples of functionality for which the cloud computing environment may be utilized. Examples of workloads and functions which may be provided from this layer include: mapping and navigation 91; software development and lifecycle management 92; virtual classroom education delivery 93; data analytics processing 94; transaction processing 95; and workload management/migration processes 96 described herein. In accordance with aspects of the invention, the workload management/migration processes 96 function operates to perform one or more of the processes described herein.
In accordance with aspects of the invention, the migration client 170 can be implemented as one or more program code in program modules 42 stored in memory as separate or combined modules. Additionally, the migration client 170 may be implemented as separate dedicated processors or a single or several processors to provide the function of these tools. While executing the computer program code, the processing unit 16 can read and/or write data to/from memory, storage system, and/or I/O interface 22. In embodiments, the program code executes the processes of the invention.
By way of example, migration client 170 may be configured to communicate with the migration server 160 via a cloud computing environment 50. As discussed with reference to
The present invention may be a system, a method, and/or a computer program product. The computer program product may include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present invention.
The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of the computer readable storage medium includes the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device such as punch-cards or raised structures in a groove having instructions recorded thereon, and any suitable combination of the foregoing. A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.
Computer readable program instructions described herein can be downloaded to respective computing/processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and/or a wireless network. The network may comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and/or edge servers. A network adapter card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing/processing device.
Computer readable program instructions for carrying out operations of the present invention may be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like, and conventional procedural programming languages, such as the “C” programming language or similar programming languages. The computer readable program instructions may execute entirely on the user’s computer, partly on the user’s computer, as a stand-alone software package, partly on the user’s computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user’s computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate arrays (FPGA), or programmable logic arrays (PLA) may execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present invention.
Aspects of the present invention are described herein with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer readable program instructions.
These computer readable program instructions may be provided to a processor of a computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks. These computer readable program instructions may also be stored in a computer readable storage medium that can direct a computer, a programmable data processing apparatus, and/or other devices to function in a particular manner, such that the computer readable storage medium having instructions stored therein comprises an article of manufacture including instructions which implement aspects of the function/act specified in the flowchart and/or block diagram block or blocks.
The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process, such that the instructions which execute on the computer, other programmable apparatus, or other device implement the functions/acts specified in the flowchart and/or block diagram block or blocks.
The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the Figures. For example, two blocks shown in succession may, in fact, be accomplished as one step, executed concurrently, substantially concurrently, in a partially or wholly temporally overlapping manner, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems that perform the specified functions or acts or carry out combinations of special purpose hardware and computer instructions.
Referring to
Neural ODEs are a family of deep neural network models. Instead of specifying a discrete sequence of hidden layers, embodiments parameterize the derivative of the hidden state using a neural network. The output of the network is computed using a differential equation solver. Neural ODEs allow end-to-end training within larger models.
Other types of neural network, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) are widely known.
Various approaches to training a machine-learning algorithm are known. Typically, such methods comprise obtaining a training dataset, comprising training input data entries 510 and corresponding training output data entries 540. An initialized machine-learning algorithm is applied to each input data entry 510 to generate predicted output data entries. An error between the predicted output data entries and corresponding training output data entries is used to modify the machine-learning algorithm. This process can be repeated until the error converges, and the predicted output data entries are sufficiently similar (e.g., ±1%) to the training output data entries. This is commonly known as a supervised learning technique.
Known methods of modifying a neural network include gradient descent, backpropagation algorithms and so on.
In embodiments, the training input data entries for the machine learning algorithm correspond to descriptions of training workloads. By way of example, a description of a workload may comprise, wherein the workload data is converted as time series of multiple variables:
In embodiments, the training output data entries for the machine learning algorithm correspond to edge locations for migrating the training workload.
In accordance with aspects of the invention, several pre-processing methods are employed to improve the training sample(s). For instance, the data sources to be used as a dataset to train the proposed machine learning algorithm may have differentiating formats. Therefore, some pre-processing may be executed to bring all training data to a common format. For this, a data pre-processor component may be used, and all data that is used to train the machine learning algorithm model may be passed through the pre-processor to achieve a consistent format of the training data.
In embodiments, the data format is designed to preserve the relationship(s) between: (i) a workload (or a set of workloads); and (ii) an edge location for migrating the workload(s). In a simple form, the data may be represented as tuples of a workload and an identifier - (workload1, edgelocationA), ([workload2, workload3], edgelocationB), etc.
In implementations, the training data comes from an initial data set (e.g., training data), or from data incoming from a workload management deployment at an edge computing environment. The initial data set may be split into a training and test data set, to provide validation for the training process.
From the above description, it will be understood that embodiments include training a machine learning algorithm based on a description of a workload (as a training input for the machine learning algorithm) and the identifier of an edge location (i.e., edge resource) as a source for migrating the workload (as a training output for the machine learning algorithm). In one example, the training comprises adjusting a configuration of the machine learning algorithm so as to decrease a difference between the identifier of the edge location for migrating the workload and a second identifier that is output by the machine learning component.
In embodiments, when a new workload is identified/obtained, a description 550 of the new workload (e.g., Z1= {V={V+V1}, A, D, T}) may be provided as an input to the artificial neural network 500. In the description of the new workload, any one of the parameter set may have a different value. A prediction result 560 may then be obtained as an output from the artificial neural network 500, wherein the prediction result 560 includes an identifier of an edge computing environment location as a target for workload migration. The identifier may then be distributed to the appropriate edge computing environment.
Embodiments include using a machine learning during a workload migration planning phase, wherein an edge layer analyzes the workloads running within the edge and “self-nominate” workloads to be migrated between edge nodes. In this manner, the workload planning machine learning models may be deployed within the edge computing environment (i.e., edge layer), because sufficient computational resources exist within the edge layer. This may help to eliminate unnecessary network traffic between the cloud and edge layers.
By way of illustration, an unmanned vehicle is an example of a “dynamic edge” application, where the number of workloads in the edge layer vary depending on the number of vehicles and their position. In this example, the machine learning model continuously optimizes the edge layer computational resources. The edge layer application software components are deployed in containers rather than VMs because containers have quicker initialization and are moved between clusters more quickly than VMs.
By way of further illustration, an exemplary use case will now be described with reference to
In this use case, first exemplary algorithm is summarized as follows:
Pseudo Code for Algorithm 1:
In the above algorithm, each workload (A) is considered a multi-variate time series. One workload A can be replicated in multiple edge locations. For each of those locations, workload A can have a different version. Workloads in edge locations are continuously changing. For example, the workloads change while the edge cluster has local cache and is constantly synchronizing with the primary workload.
In embodiments, to compute the changes, boundary condition bc may be considered. For example, the current version at time t for each i-th location is included in the model training. In one example, no workload older than τ is included in the training for the machine learning model. In examples, to cater for significant and continuous changes, a Neural Ordinary differential equation (ODE) solver is employed to determine the changes in workload for small time changes. The Neural ODE solver takes a Distributed Neural Network (DNN) model and trains it with the workload controlled by boundary conditions (bc).
In embodiments, the next step is to determine the suitable edge location(s) that can be used for optimized migration of workload A with the best achievable availability characteristics (e.g., minimal downtime through migration). The busiest version of workload A may have more impact than another replica of A at different edge location. On the other hand, the latest version of A may obtain maximum migration from single point.
In embodiments, the machine learning algorithm seeks to select the most optimum edge location as the source of migration of workload A. For this, each edge location is ranked based on their frequency of changes and access δt. From line 5 to line 10 in Algorithm 1 above, the exemplary training method selects the suitable edge location for workload A migration based on latest version of workload A as well as mostly accessed or busiest edge location for A.
So that the proposed learning model may learn the business impact for each edge locations, embodiments may employ a loss function for the gradient of the neural network model. Instead of conventional democratic selection of edge location, embodiments compute the business impact (δt) of each edge location. Edge locations with lowest business impact may then be ranked higher as suitable candidates for migration. At the same time, the latest versions of workloads are ranked higher as suitable candidate for migration source location. To balance these two opposite controlling factors, a loss function may be defined which also controls the gradient descents of the proposed distributed neural network model. An exemplary algorithm for this may be summarized as follows:
Algorithm 2 - loss_fn(δAi, δt)
Pseudo Code for Algorithm 2:
Instead of migrating any workload from a single source node for T time making RTO= T, embodiments may choose multiple locations for workload migration in parallel. By way of example, an exemplary algorithm for parallel migration from multiple edge locations may be summarized as follows:
Pseudo Code for Algorithm 3:
Furthermore, embodiments may also be configured to adapt/improve based on monitoring and/or feedback data. For instance, to verify how well the machine learning algorithm/model performs when suggesting edge locations, the migration metrics may be tracked. This could be monitored either on a per-edge device basis, or throughout all deployments across a network environment. During and/or after migration, the edge device may verify and/or assess the migration by analyzing activity logs.
Collected migration data may be made accessible by an operator or a Site Reliability Engineer, so that manual analysis and/or modifications may be made.
It should now be understood by those of skill in the art, in embodiments of the present invention, the proposed machine-learning concepts provide numerous advantages over conventional workload migration approaches. These advantages may include, but are not limited to:
In still further advantages to a technical problem, the systems and processes described herein provide a computer-implemented method for efficient workload management and/or workload migration at edge computing environments. In this case, a computer infrastructure, such as the computer system shown in
The descriptions of the various embodiments of the present invention have been presented for purposes of illustration, but are not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The terminology used herein was chosen to best explain the principles of the embodiments, the practical application or technical improvement over technologies found in the marketplace, or to enable others of ordinary skill in the art to understand the embodiments disclosed herein.
