Computer implementation methods, systems, and computer program products (interpolating performance data)

Abstract

We provide computer implementation methods, systems, and computer program products for interpolating performance data. [Solution] The method includes determining an event occurring in a first time related to a computing system; acquiring system data related to the computing system; determining the system state of the computing system in the first time based on the system data; determining two or more system data clusters containing clustered system data related to the system state of the computing system based on the system state; determining an interpolated data value for the first time based on the system data via an interpolation algorithm; and adjusting the interpolated data value based on the determination that the interpolated data value lies outside of the two or more system data clusters.

Description

Technical Field

[0001] The present invention generally relates to data management, and more specifically, to an improved method for interpolating performance data in a computing system. and is related to the relevant performance data interpolation method in the computing system.

Background Art

[0002] The performance of a computer system changes during its execution. Therefore, a certain system may exhibit good performance in some parts during execution and may experience performance degradation at other times. In order to understand why the system behaves in such a way, performance analysts study the behavior over time. Most recent microprocessors support a hardware performance monitor that can collect performance metrics that vary temporally during the execution of a computer system. However, although these processors can collect hundreds of performance metrics, the number that can be collected simultaneously is limited, and they can only be collected during established sampling times. Due to such limitations in the collection of performance data, various interpolation techniques can be used to determine performance metrics during periods when performance data of the computer system was not collected. Modern approaches attempt to interpolate performance metrics using interpolation techniques such as arithmetic mean and linear regression, but they are often inaccurate, for example, due to considering only sampled performance data. but the number that can be collected simultaneously is limited, and they can only be collected during established sampling times. Because there are such limitations in the collection of performance data, various interpolation techniques can be used to determine performance metrics during periods when performance data of the computer system was not collected. Modern approaches attempt to interpolate performance metrics using interpolation techniques such as arithmetic mean and linear regression, but they are often inaccurate, for example, due to considering only sampled performance data. to determine performance metrics during periods when performance data of the computer system was not collected. Modern approaches attempt to interpolate performance metrics using interpolation techniques such as arithmetic mean and linear regression, but they are often inaccurate, for example, due to considering only sampled performance data. to determine performance metrics during periods when performance data of the computer system was not collected. Modern approaches attempt to interpolate performance metrics using interpolation techniques such as arithmetic mean and linear regression, but they are often inaccurate, for example, due to considering only sampled performance data. to interpolate performance metrics, but are often inaccurate, for example, due to reasons such as considering only sampled performance data. which is often inaccurate due to reasons such as considering only sampled performance data.

Summary of the Invention

Problems to be Solved by the Invention

[0003] Provide a computer-implemented method, system, and computer program product for interpolating performance data. to provide a computer-implemented method, system, and computer program product for interpolating performance data. [Means for solving the problem]

[0004] Embodiments of the present invention involve interpolating performance data in a computing system. It is directed towards. Non-exclusive illustrative computer implementation methods are computing systems The purpose is to determine the events related to the stem, where the events occur in the first time period. Making decisions and obtaining system data related to the computing system. Based on the system data, the system of the computing system in the first time period Determining the system state and, based on the system state, the computing system Two or more system data clusters containing system data related to the system state The first raster is determined based on the system data via an interpolation algorithm. Determining the interpolated data value for time, and the interpolated data value for two or more system days This includes adjusting interpolated data values ​​based on the decision that they are outside the cluster.

[0005] Embodiments of the present invention involve interpolating performance data in a computing system. It is directed towards. Non-exclusive illustrative computer implementation methods are computing systems This involves determining system data related to the system, and the system data is a multiple type Includes multiple performance metrics related to the computing system obtained by ImSlice. Furthermore, using a neural network model, we can analyze the performance of each of the multiple performance metrics. Generating strength values ​​and directly applying performance robustness values ​​to each of multiple performance metrics. Plotting on the intersection plane and the performance robustness value of the computing system At the very least, clustering into one system state, and at least one system state This involves generating a data cluster for each state, and the data cluster is a computer This includes generating and relating to performance metrics of the tuning system.

[0006] Another embodiment of the present invention features the above-described method in a computer system and a computer This is to be implemented in the data program product.

[0007] Additional technical features and advantages are realized by the technology of the present invention. Embodiments of the present invention And aspects thereof are described in detail herein and are considered part of the claimed subject matter. For a better understanding, please refer to the detailed explanations and drawings. [Brief explanation of the drawing]

[0008] [Figure 1] This figure shows an exemplary graph illustrating current interpolation techniques for estimating performance metrics in computing systems. [Figure 2] This figure shows a system for interpolating performance index values ​​in a computing system according to one or more embodiments of the present invention. [Figure 3] This figure shows exemplary performance metrics collected during various time slices according to one or more embodiments of the present invention. [Figure 4] This figure shows exemplary clustering of performance data on an orthogonal plane according to one or more embodiments of the present invention. [Figure 5] This figure shows exemplary clusters for each performance category according to one or more embodiments of the present invention. [Figure 6] This figure shows a plot of interpolated data values ​​in a one-dimensional array having data clusters according to one or more embodiments of the present invention. [Figure 7]A diagram showing an exemplary methodology for determining which data clusters adjust interpolation values according to one or more embodiments. [Figure 8] A diagram showing two scenarios regarding the arrangement of interpolation values within a data cluster according to one or more embodiments of the present invention. [Figure 9] A diagram showing a flowchart of method 900 for interpolating performance data in a computing system according to one or more embodiments of the present invention. [Figure 10] A diagram showing a flowchart of method 1000 for interpolating performance data in a computing system according to one or more embodiments of the present invention. [Figure 11] A diagram showing a cloud computing environment according to one or more embodiments of the present invention. [Figure 12] A diagram showing an abstraction model layer according to one or more embodiments of the present invention. [Figure 13] A diagram schematically showing a computer system according to an embodiment.

Mode for Carrying Out the Invention

[0009] The specific content of the exclusive rights described in this specification is particularly pointed out and clearly claimed in the claims at the conclusion of the specification. The foregoing and other features and advantages of the embodiments of the present invention will be apparent from the following detailed description taken in conjunction with the accompanying drawings.

[0010] The diagrams depicted in this specification are illustrative. Without departing from the spirit of the present invention, there can be many variations to the diagrams or the operations described therein. For example, the operations can be executed in a different order, or the operations can be added, deleted, or modified. Also, the term "coupled" and its variations are used to describe having a communication path between two elements, and This does not mean that elements are directly connected without any intermediary elements or connections between them. All of these variations are considered part of this specification.

[0011] One or more embodiments of the present invention perform interpolation using a computing system By capturing the state of the system, it improves the interpolation of computing system data. The system and method are provided. The performance of a computer system changes during its execution. Therefore, it is necessary to collect performance data to track the performance metrics needed to address the problem. Performance data collection typically involves sampling data according to a defined sampling rate. This is done with hardware registers. However, these processors have hundreds of performance levels. Although it is possible to collect indicators, the number that can be collected simultaneously is limited, and established standards Data can only be collected during the pull-up rate. This presents a limitation in the collection of performance data. Therefore, performance indicators for periods when performance data for the computing system was not collected are not available. Various interpolation techniques are used to determine the target.

[0012] Figure 1 shows current interpolation techniques for estimating performance metrics in computing systems. An exemplary graph illustrating the technique is shown. Graph 102 shows how performance indicator data is used. This indicates whether the pump is working, and is shown as a graph with performance indicators on the Y-axis and time on the X-axis. As shown in Graph 102, sampling of this performance metric is performed every 5 minutes. In some cases, in a computing system, sampled performance metrics are An event occurs that doesn't exist, and the computing engineer wants to determine performance metrics. This can happen. In the illustrated example, two data points are found at 1:00 AM and 1:05 AM. The collected event occurred at 1:02 AM. The second graph, 104, shows the event at 1:02 AM. Two interpolation techniques (arithmetic mean and linear regression) used to interpolate performance metrics are shown. As shown in Graph 104, the two interpolation techniques produce two results that are based on actual performance data. It is much higher than the result of the previous result, so it produces inaccurate results.

[0013] As mentioned above, current interpolation techniques include arithmetic mean, linear regression, or a combination thereof. There is a reason for this. In arithmetic mean, two performance indicator values ​​are used in the process. Also, in linear regression, 4 Multiple performance metrics are used in the processing. However, these methods belong to the same category (same performance). Only performance metric data belonging to the metric category will be used. For example, network input / output (IO). When predicting bandwidth, only historical I / O bandwidth data is used for interpolation. When implementing with Nario, network I / O bandwidth depends on CPU utilization, task priority, and data Various factors such as the disk I / O status, the number of TCP connections, or other factors or combinations thereof. Factors that may influence this include, but are not limited to, these other categories. By analyzing the data, we can improve the interpolation of performance data.

[0014] Now, looking at an overview of the embodiments of the present invention, one or more embodiments of the present invention are: To provide an interpolation methodology that takes into account the state of the computing system when performing interpolation. This addresses the aforementioned shortcomings of prior art. The state of the computing system is Lanced resources, memory depletion, high CPU, memory, and I / O utilization, network This includes, but is not limited to, network I / O contention and storage I / O contention. It is not that. Using historical performance metric data and clustering techniques, the computer In a ting system, the system state is determined by various data clusters. This is possible. Therefore, when interpolating performance indicator values, the state of the computing system Determine this and check or adjust the interpolation performance index values ​​of these data clusters, or both. It is possible to improve accuracy by using this method.

[0015] Figure 2 shows a computing system according to one or more embodiments of the present invention. This diagram shows a system for interpolating performance indicator values. System 200 is a computer. Configure to receive sampled performance data 204 for the running system. The system includes a controller 202. The system 200 also includes a controller 202 that performs performance data Sampled, which enables outputting enhanced interpolation value 212 for data 204. The performance data 204 can be used by the controller 202 to process it. Includes a clustering engine 206. In one or more embodiments of the present invention, Sun The pulled performance data 204 can be received by the controller 202. The controller 202 uses an interpolation algorithm to interpolate values ​​for a specific time period. It can be generated. The interpolation algorithm can be, for example, an arithmetic mean or a linear regression algorithm. It can be rhythm or both. The specific time for interpolation is determined by the performance data collected. It can be determined based on events occurring in the computing system. The event could be, for example, a performance problem that occurred at a certain time. To address this, a snapshot of the performance metrics at that point in time is needed, therefore this snapshot For a close-up shot, it's necessary to interpolate performance metrics at the time of the event. As mentioned above... Therefore, instead of just using the interpolation algorithm, the controller 202 also uses interpolated values. To further adjust / enhance the computing system, analyze its state. By establishing the state of the interpolation system, the controller 202 adjusts the interpolation value. When deciding whether or not to do so, you can look up historical performance data values ​​for specific performance metrics. The historical performance data values ​​are obtained using clustering engine 206, and the data cluster These data clusters can be arranged to ensure that the interpolation performance values ​​are accurate. Alternatively, comparison metrics are considered to determine whether adjustments are necessary. For example, interpolation performance. If the values ​​are within a data cluster for the system state, interpolation is accurate and aligns with that. It probably won't be necessary. However, if the interpolation performance value is outside the data cluster, the interpolation performance value It will be necessary to adjust this. In this case, move the value into the data cluster and control This can be returned as an enhanced interpolation value of 212.

[0016] In one or more embodiments of the present invention, the clustering engine 206 is one or It utilizes multiple machine learning algorithms, neural networks, or both. Then, performance data is clustered. All available performance data of the target system Performance data for categories (metrics) can be collected, and the computing system To quantify the stem state, it can be used to train machine learning models. Figure 3 shows the data collected during various time slices according to one or more embodiments of the present invention. This figure shows exemplary performance metrics. In the illustrated example, the collected performance metrics are C PU utilization, memory usage, system disk read / write bytes per second Number, network bandwidth (intranet), and network bandwidth (intra This includes, but is not limited to, net-out. These performance metrics data are Mapped to a specific system state and recorded for use in the interpolation method described herein. It is possible to do so. The data mass cluster determines the performance of each system state. It is possible to generate data for categories.

[0017] In one or more embodiments of the present invention, all time slice performance data are The performance states on the Cartesian coordinate system can be clustered. Figure 4 shows one or more of the present invention. This document illustrates exemplary clustering of performance data in an orthogonal plane according to multiple embodiments. There is. The orthogonal plane 400 is, for example, the Y-axis which shows the performance of the computing system, and It can have an X-axis that indicates the robustness of the computing system. The plotted performance metrics are based on the performance robustness of the neural network model. Weights can be applied to create performance values. The performance and robustness of the input performance values ​​are also considered. The s can be calculated using the Radial Basis Function (RBF) model. The feedforward network includes an input layer, a hidden layer, and an output layer. The RBF network ensures that the data points of each system are efficient in terms of performance and robustness. Generate a value within a range that indicates how low or how high the impact on both is. For example, a data point in a certain system may have little impact on performance, but its robustness The impact is significant. On the other hand, data points from another system have a shadow on performance. It's possible that the resonance is high, but the impact on robustness is low. The data is plotted on these orthogonal coordinates, and the performance is calculated using the mean shift algorithm. The system is clustered into states (S1~S6). Several exemplary performance states S1: Highly utilized CPU, memory, and I / O; S2: Balanced resources. S3: Memory exhaustion, S3: CPU contention, S5: Network I / O contention, and S6: Includes storage I / O contention. These descriptors are illustrative and do not limit the types of performance states. This was not intended. Furthermore, different computing systems may have different states. It can happen.

[0018] In one or more embodiments of the present invention, the performance data mass cluster comprises each performance state This is created for each performance category. Raw performance data and performance data in an orthogonal plane. When a mapping to the strength data is recorded, a specific system state (e.g., S1) is recorded. Past performance data can be obtained for ). For S1, the CPU utilization is The collected data is placed in a one-dimensional array. The original data values ​​are clustered using the mean. A shift algorithm is used. This technique does not require prior knowledge of the number of clusters. This is a parametric clustering technique. Figure 5 shows one or more embodiments of the present invention. This figure shows exemplary clusters for each performance category in each system state S.k to Then, the performance values ​​for each category are processed, and a mass data cluster like the one shown in Figure 5 is created. For each cluster, the centroid value is calculated and the interpolation performance data values ​​are adjusted. It is used to determine whether or not, and how to adjust it.

[0019] In one or more embodiments of the present invention, the controller 202 first Using the available data of the computing system in the time slice, Interpolates performance index values. The system state for the computing system is determined, and As described above, data clusters related to the system state are obtained. At this point, the first time For the system state in between, interpolation performance values ​​are calculated based on these data clusters. A decision is made on whether or not to adjust it. Figure 6 shows one or more embodiments of the present invention. The figure shows a plot of interpolated data values ​​in a one-dimensional array with data clusters. Yes, there are two examples of one-dimensional arrays. Example 602 shows that the interpolated values ​​are data clusters (clusters). This shows an example where the interpolation value does not need to be adjusted, within the 3) range. Based on the state, and according to the historical data used to create these clusters, The interval indicates that the exact value was returned. In case 604, the interpolated value was either a data cluster in the array or There are examples of interpolation values ​​that are out of line and therefore require adjustment. If it is determined that the data is outside the data cluster, the interpolated value will be based on the existing data cluster. It is adjusted. Figure 7 shows which data clusters are interpolated in one or more embodiments. This diagram illustrates an exemplary methodology for determining whether to adjust the values. The array shows the interpolated values. This includes clusters 2 and 3 located between the two data clusters. Exemplary calculation This includes gravity calculations, including equation [1]. TIFF2026090434000002.tif9134

[0020] In equation [1], F Gravity_cluster Gravity from the cluster, N cl uster `dis_cluster` is the number of points in the cluster, and `dis_cluster` is the cluster centroid and complement. This is the distance from the intermediate value. In Figure 7, the initial interpolated value is located between cluster 2 and cluster 3, for example. The gravity value is calculated using equation [1]. The result of the calculation is that the gravity of cluster 3 is equal to that of cluster 2. This becomes greater than gravity, indicating that the interpolated values ​​are drawn into cluster 3, as shown in the diagram. This is based on gravity calculations using the centroids of the data clusters, and interpolates the values ​​to determine which cluster This represents the decision to draw them into the cluster, and is sometimes called the "nearest" cluster. Once the cluster is determined, the next decision is where to place the interpolated values ​​within the "nearest" cluster. This will determine whether or not. Figure 8 shows one or more embodiments of the present invention. Two scenarios are shown for the placement of interpolated values ​​within a cluster. In Scenario 802... This calculates the midpoint between the interpolated value and the cluster centroid. The midpoint is the data class Since it is within cluster 3, the interpolated value is adjusted / modified to be the value of the midpoint, and It is returned as an interpolated value. In scenario 804, the midpoint is still outside the data cluster. In this scenario, the interpolated values ​​are adjusted / modified to become the cluster boundary values, and the boundary values ​​are It is returned as an enhanced interpolation value.

[0021] In one or more embodiments of the present invention, the controller 202 and the system 20 Any of the components on 0 are implemented on the processing system 1300 shown in Figure 13. It is possible. Furthermore, the cloud computing system 50 is system 200 Cloud50 can communicate with one or all of its elements via wired or wireless electronic communication. , complementing, supporting, or replacing some or all of the functions of the elements of System 200 It is possible. Furthermore, some or all of the functions of the elements of System 200 are Cloud 5 It can be implemented as node 10 of 0 (shown in Figures 11 and 12). Cloud computing node 10 is an example of a suitable cloud computing node. This does not imply any limitations on the scope of use or functionality of the embodiments of the present invention described herein. It is not intended to do so.

[0022] Figure 9 shows a computing system according to one or more embodiments of the present invention. A flowchart of Method 900 for interpolating performance data is shown. Some of these are executed by, for example, one or more processors 1301 from Figure 13. This can be done. Method 900 is a computing system as shown in block 902. This includes determining the event related to the first time, which will occur. In lock 904, method 900 relates to a system related to a computing system. This includes obtaining data. System data is obtained in block 906 by method 900. However, this includes determining the system state of the computing system in the first time period. Next, the system data is determined. Also, in block 908, method 900 determines the system Based on the state, clustering related to the system state of the computing system This includes determining two or more system data clusters containing the selected system data. Furthermore, method 900, as shown in block 910, uses an interpolation algorithm to enable the system This includes determining interpolated data values ​​for a first time based on the data. In lock 912, method 900 is a system data class in which the interpolated data values ​​are two or more system data classes. This includes adjusting the interpolation value based on the decision that it is outside the range.

[0023] Additional steps may also be included. The steps shown in Figure 9 are illustrative and are not within the scope of this disclosure. And without deviating from the spirit, other processes may be added, and existing processes may be deleted or modified. Alternatively, please understand that rearrangement is permitted.

[0024] Figure 10 shows a computing system in one or more embodiments of the present invention. A flowchart of Method 1000 for interpolating performance data is shown. At the very least, some of them are, for example, by one or more processors 1301 from Figure 13. It can be executed. Method 1000 is a computer as shown in block 1002. This includes determining system data related to the tuning system, and the system data is Multiple performance indicators related to the computing system obtained across multiple time slices Includes reference values. In block 1004, method 1000 is a neural network model. This includes using a tool to generate performance robustness values ​​for each of multiple performance metrics. Furthermore, in block 1006, method 1000 provides for each of the multiple performance indicator values. This includes plotting the performance robustness values ​​on an orthogonal plane. Block 1008 Method 1000 is a method that sets the performance robustness value of a computing system to at least 1 This includes clustering into two system states. And in block 1010 Method 1000 generates a data cluster for each of at least one system states. Including this, data clusters are associated with performance metrics of computing systems. It is possible.

[0025] Additional steps may also be included. The steps shown in Figure 10 are illustrative and are not within the scope of this disclosure. Other processes may be added, and existing processes may be deleted or modified, without deviating from the framework and spirit. Please understand that you may rearrange or reposition them.

[0026] This disclosure includes a detailed description of cloud computing, but is not described herein. It was understood that the implementation of the teachings provided is not limited to cloud computing environments. Rather, embodiments of the present invention are not related to any other types currently known or to be developed later. It can be implemented in combination with the P computing environment.

[0027] Cloud computing is a shared pool of configurable computing resources. (For example, network, network bandwidth, server, processing, memory, storage device, app) Convenient and on-demand network access to applications, virtual machines, and services. This is a service delivery model that enables the use of services with minimal management effort or minimal management effort. Through communication with the service provider, we can quickly prepare (provision) and release the service. This cloud model has at least five characteristics and at least three servers. It may include a visor model and at least four deployment models.

[0028] The characteristics are as follows:

[0029] On-demand self-service: Cloud consumers can avoid human interaction with service providers. Without requiring any interaction, it automatically adjusts server time and network storage as needed. This allows for the unilateral provision of computing capabilities such as those mentioned above.

[0030] Broad network access: Computing power is available via the network. It is possible and accessible through standard mechanisms. This allows heterogeneous systems to access each other. or thick client platform (e.g., mobile phones, laptops, PDs) Use by A) will be promoted.

[0031] Resource pooling: The provider's computing resources are pooled, multi It is offered to multiple consumers using a tenant model. Various physical and virtual resources Sources are dynamically allocated and reallocated according to demand. Generally, consumers are provided Because the exact location of the resources is not managed or known, location-independent (location i There is a sense of independence. However, consumers have a higher level of abstraction (e.g., country, state, In some cases, the location can be identified (in data centers).

[0032] Rapid flexibility: Computing power should be able to be prepared quickly and flexibly. Because it can do this, in some cases it can automatically and immediately scale out, and also quickly re It can be leased and immediately scaled in. For consumers, it is available for preparation. Computing power often appears unlimited, and you can purchase any quantity at any time. It is possible.

[0033] Services measured: Cloud systems are defined by the type of service (e.g., storage, processing). A measuring instrument with a certain level of abstraction suitable for (process, bandwidth, active user accounts) Leverage the capabilities to automatically control and optimize resource usage. Monitor resource usage and control We will report and provide transparency to both the service providers and consumers using the services. It can be provided.

[0034] The service model is as follows:

[0035] Software as a Service (SaaS): The functions provided to consumers are cloud-based. This means that you can use the provider's applications that run on the infrastructure. The application in question is a thin client such as a web browser (e.g., webmail). It can be accessed from various client devices via the user interface. This includes networks, servers, operating systems, storage, and individual applications. The management and control of the underlying cloud infrastructure, including even the communication functions, are performed. No. However, this does not apply to user-specific, limited application configuration settings. .

[0036] Platform as a Service (PaaS): The functions provided to consumers are provided by PaaS. Using programming languages ​​and tools supported by IDA, consumers can create or Deploying the acquired application to the cloud infrastructure. This includes the consumer's network, servers, operating system, and storage. Hmm, we do not manage or control the underlying cloud infrastructure, but the deployed A It allows you to control the application and, in some cases, the configuration of its hosting environment. ru.

[0037] Infrastructure as a Service (IaaS): The functions provided to consumers are Consumers can deploy any software, including operating systems and applications. and executable processor, storage, network, and other basic computing The task is to prepare the necessary resources. Consumers need to prepare the underlying cloud infrastructure. It does not manage or control the structuring, but it does manage the operating system, storage, and development. Open applications can be controlled, and in some cases, some network components It allows for partial control over the network (e.g., the host firewall).

[0038] The deployment model is as follows:

[0039] Private Cloud: This cloud infrastructure is operated exclusively for a specific organization. This cloud infrastructure is managed by the organization or a third party. It can exist on-premises or off-premises.

[0040] Community Cloud: This cloud infrastructure is shared by multiple organizations. They have common interests (e.g., mission, security requirements, policies, and cons). Supports specific communities that have policies. This cloud infrastructure Kucha can be managed by the organization or a third party, on-premises or off-premises. It can exist as a premise.

[0041] Public Cloud: This cloud infrastructure is used by a large number of people and large organizations. It is provided to a model industry association and owned by an organization that sells cloud services.

[0042] Hybrid Cloud: This cloud infrastructure consists of two or more cloud models. It will be a combination of Dell (private, community, or public). Each model retains its own unique entity, but is bound by standards or individual technologies. Data and application portability (e.g., cloud bars for load balancing between clouds) To achieve sting.

[0043] Cloud computing environments are characterized by statelessness and low coupling. (low coupling), modularity, and semantic interoperability It is a service-oriented environment that emphasizes low-cost performance. At its core is an infrastructure that includes a network of interconnected nodes. be.

[0044] Here, Figure 11 shows an example of a cloud computing environment 50. In addition, the cloud computing environment 50 is one or more cloud computing nodes This includes D10. In contrast to these, the local computer devices used by cloud consumers ( For example, PDA or mobile phone 54A, desktop computer 54B, laptop Automotive computer 54C, or automotive computer system 54N, or combination thereof. Nodes (such as those linked together) can communicate with each other. Nodes 10 can communicate with each other. Code 10, for example, refers to the aforementioned private, community, public, or hybrid In one or more networks, such as a cloud or a combination thereof, Alternatively, they can be virtually grouped (not shown). This allows cloud computing The working environment 50 is infrastructure as a service, platform and They can provide either software or a combination thereof, and cloud consumers can Regarding these, there is no need to maintain resources on the local computer device. The types of computer devices 54A to N shown in 11 are merely examples, and computing The 10 and cloud computing environments 50 can use any type of network or This refers to a network addressable connection (e.g., using a web browser) or both. Please understand that it is possible to communicate with any type of electronic device via this method.

[0045] Here, the functional extraction provided by the cloud computing environment 50 (Figure 11) The set of embodied layers is shown in Figure 12. Note that the components, layers and The functions are merely illustrative, and the embodiments of the present invention are not limited to these. As shown in the diagram, the following layers and corresponding functions are provided.

[0046] Hardware and software layers 60 are hardware components and software Includes software components. Examples of hardware components include mainframes. M61, a server based on a reduced instruction set computer (RISC) architecture 62, Server 63, Blade Server 64, Storage Device 65, and Network and Network A software component 66 is included. In some embodiments, the software component The network application server software 67 and database Includes software 68.

[0047] The virtualization layer 70 provides an abstraction layer. From this layer, for example, the following virtual elements It can provide: virtual server 71, virtual storage 72, virtual private Virtual networks including virtual networks, virtual applications and operators A ting system 74, and a virtual client 75.

[0048] As an example, the management layer 80 can provide the following functions: Resource preparation 81 This refers to computers used to perform tasks within a cloud computing environment. Enables dynamic procurement of tung resources and other resources. Quantitative and pricing 8 2 is cost tracking when resources are used within a cloud computing environment, and This enables billing or invoice issuance for the consumption of these resources. These resources may include application software licenses. The company not only protects data and other resources, but also supports cloud consumers and It enables the identification and verification of the user portal 83. Provides access to the loud computing environment. Service level management 84 is required. To ensure that the required service level is met, cloud computing resources are allocated accordingly. To enable allocation and management. Service Level Agreement (SLA) planning and execution 85. In accordance with the SLA, advance provision of cloud computing resources that are expected to be needed in the future. It enables arrangement and procurement.

[0049] Workload Layer 90 is an example of the capabilities available in a cloud computing environment. It provides. Examples of workloads and functions that can be provided from this layer include mapping and Navigation 91, Software Development and Lifecycle Management 92, Virtual Classroom Education Distribution 93, data analysis processing 94, transaction processing 95, computing system performance The interpolation of 96 is included.

[0050] Here, in Figure 13, the computer system 1300 in the embodiment is generally shown. The computer system 1300 is configured to support various communication technologies as described herein. Any number and combination of computing devices and networks that utilize the technology To include, utilize, or both, an electronic or computer framework. This is possible. Computer system 1300 can change to different services, and several It has the ability to independently reconfigure functions, is easily scalable, and is expandable. , and may be modular. The computer system 1300 is, for example, Desktop computers, laptop computers, tablet computers , or a smartphone. In some examples, a computer system 130 0 may be a cloud computing node. Computer system 130 0 is a computer system executable instruction such as a program module. It may also be explained in the general context of being executed by a system. Generally, a program Modules are routines that perform a specific task or implement a specific abstract data type. This may include programs, objects, components, logic, data structures, etc. i. The computer system 1300 is linked via a communication network to remote In a distributed cloud computing environment where tasks are performed by processing devices It may be implemented in a distributed cloud computing environment. The system includes both local and remote computer system storage media, including memory storage devices. It may be placed in [location].

[0051] As shown in Figure 13, the computer system 1300 is one or more central processing units (CPU) 1301a, 1301b, 1301c, etc. (general term or generally referred to as processor) It has a single-core processor (referred to as 1301). The processor 1301 is a single-core processor, multi-core It can be an a-processor, a computing cluster, or any number of other configurations. The processor 1301, also called the processing circuit, can communicate via the system bus 1302. It is coupled to system memory 1303 and various other components. System memory 13 03 consists of read-only memory (ROM) 1304 and random access memory (RAM). 1305 may be included. ROM 1304 is coupled to system bus 1302 and Basic Input / Output System (BI) controls specific basic functions of the Computer System 1300. It can include the OS. RAM is used by the system battery of processor 1301. This is a read / write memory coupled to S1302. System memory 1303 is a moving The system provides a temporary memory space for the execution of the aforementioned instruction. System memory 1303 These include random access memory (RAM), read-only memory, flash memory, and Alternatively, it may include any other suitable memory system.

[0052] The computer system 1300 has inputs / outputs (I) coupled to the system bus 1302. I / O adapter 1 consists of adapter 1306 and communication adapter 1307. 306 is the hard disk 1308 or any other similar component or A small computer system interface (SCSI) adapter that combines and communicates with It is possible. The I / O adapter 1306 and the hard disk 1308 are specified in this specification. Collectively, this is called mass storage 1310.

[0053] Software 1311 for running on computer system 1300 is mass-produced. It may be stored in rage 1310. Mass storage 1310 is connected to processor 1301. Therefore, it is an example of a readable tangible storage medium, and the software 1311 is a computer Instructions to be executed by processor 1301 to operate system 1300 It is stored as follows, for example, with regard to various figures as described below. Examples of computer program products and the execution of such instructions are described in more detail herein. This is explained below. The communication adapter 1307 uses the system bus 1302 to connect to an external network. The computer system 1300 may interconnect with network 1312, and other This enables communication with such a system. In one embodiment, system memory 13 Part of 03 and the mass storage 1310 function as various components shown in Figure 13. To coordinate, IBM Corporation's z / OS or AIX operators Any suitable operating system such as a running system, etc. It stores the entire system together.

[0054] Additional input / output devices include the display adapter 1315 and interface adapter. It is shown to be connected to system bus 1302 via TA 1316. In this configuration, adapters 1306, 1307, 1315, and 1316 are intermediate bus bridges (as shown in the diagram). Connected to one or more I / O buses connected to system bus 1302 via (not) It may be. Display 1319 (e.g., screen or display monitor) It is connected to the system bus 1302 by the display adapter 1315, which is, Graphics control to improve performance in graphics-intensive applications It may also include a ra and video controller. Keyboard 1321, mouse 1322, Speakers 1323 and others connect to system bus 13 via interface adapter 1316. 02 can be interconnected, which means, for example, multiple device adapters can be integrated into a single unit. The circuit may include a super I / O chip. (Hard disk controller) For connecting peripheral devices such as network adapters and graphics adapters. A suitable I / O bus is typically a common protocol such as Peripheral Components Interconnect (PCI). It includes. Therefore, as configured in Figure 13, the computer system 1300 is Processing capacity in the form of processor 1301, and system memory 1303 and Storage capacity including storage 1310, keyboard 1321 and mouse 1322 Input means such as, and output capabilities including speaker 1323 and display 1319 It contains power.

[0055] In some embodiments, the communication adapter 1307 is particularly useful for small internet connections. Use any suitable interface or protocol, such as a computer system interface. It can be used to transmit data. Network 1312 is, in particular, a cellular network. Work, wireless network, wide area network (WAN), local area network It may be a network (LAN) or the internet. External computing data The vice may connect to the computer system 1300 via network 1312. In some cases, an external computing device may be an external web server or It could also be a cloud computing node.

[0056] The block diagram in Figure 13 shows the computer system 1300 as shown in Figure 13. Please understand that this is not intended to indicate that it includes all of the elements. Rather, it is intended to indicate that the compilation The computer system 1300 may have any suitable few or additional options not shown in Figure 13. Components (e.g., additional memory components, embedded controllers, modules) It can include (a wire, additional network interfaces, etc.). With respect to the data system 1300, the embodiments described herein may be any appropriate logic It may be implemented in any way, and the logic referred to herein may be optional in various embodiments. Appropriate hardware (e.g., processor, embedded controller, or application) (Specific integrated circuits, etc.), software (e.g., applications), firmware Any suitable combination of wear, or hardware, software, and firmware. It is possible to include "se".

[0057] Various embodiments of the present invention are described herein with reference to the relevant drawings. Alternative embodiments of the present invention can be devised without departing from the scope of the present invention. The description and drawings show various connections and positional relationships between elements (e.g., above, below, adjacent). (etc.) are shown. These connections, or their positional relationships, or both, are not specifically indicated. To the extent that it is direct or indirect, the present invention is intended to be limited in this respect. It is not that. Therefore, the joining of entities is either a direct or indirect joining. It can point to something, and the positional relationship between entities can be direct or indirect. Furthermore, the various tasks and process steps described herein are provided in this specification. Having additional steps or features not described in detail would make the procedure more comprehensive. It can be incorporated into the process.

[0058] One or more of the methods described herein are any of the following techniques well known in the art. Alternatively, it can be implemented in combination. Logic for implementing logical functions in data signals. Inconspicuous logic circuits with gates, application-specific circuits with appropriate combinations of logic gates Integrated circuits (ASICs), programmable gate arrays (PGAs), field programming Examples include multi-gate arrays (FPGAs).

[0059] For brevity, prior art relating to creating and using aspects of the present invention is: Various technical aspects may or may not be described in detail herein. In particular, various technical aspects described herein Computing systems and specific computer programs for implementing features Various aspects of this are well known. Therefore, for brevity, many of the details of conventional implementations will be omitted. or a well-known system or process that is merely briefly mentioned herein. or both details are completely omitted without being provided.

[0060] In some embodiments, various functions or operations are performed at a given location, or one of them. This can be done in relation to the operation of multiple devices or systems, or both. In some embodiments, a given function or operation is performed on a first device or location. It can be run on one or more additional devices or It can be done in a specific location.

[0061] The terms used herein are for the sole purpose of describing specific embodiments. It is not intended to be limiting. When used herein, the singular "a", "a The words "n" and "the" can also refer to the plural unless the context clearly indicates otherwise. The term "comprises" or "includes" is used herein. The term "comprising" or both refers to the listed features, integers, steps, Specifies the existence of an operation, element, or component, or a combination thereof, but only one or more. Other features, integers, steps, operations, elements, components, or groups of them above This does not preclude the existence or addition of P or any combination thereof.

[0062] All means or step-plus-function pairs in the following claims The corresponding structure, materials, operation, and equivalents are combined with elements of other claims specifically claimed. It is intended to include any structure, material, or action necessary to perform the function in conjunction with it. The description of the present invention is presented for illustrative and explanatory purposes, but is not exhaustive or limited. The present invention is not intended to be limited to the embodiments shown. Many modifications and variations are possible. Without departing from the scope and spirit of the invention, it will be obvious to those skilled in the art. Embodiments are as follows: To best explain the principles and practical applications of the present invention, and to those skilled in the art, the specifics intended are described below. To help you understand the present invention in various embodiments with various modifications to suit the application These are the ones that were selected and explained.

[0063] The figures depicted herein are illustrative. Without departing from the spirit of this disclosure, Many variations are possible with respect to the steps (or actions) described in the figures or specification. For example, actions can be performed in different orders, or actions can be added, deleted, or modified. It is possible. Furthermore, the term "coupled" refers to having a signal path between two elements. This describes the process and signifies a direct connection between elements without any intermediary elements or connections. These are not modifications. All of these modifications are considered part of this disclosure.

[0064] The following definitions and abbreviations are used in the interpretation of the claims and specification. When used in this context, it means "comprises", "comprising", or "include". s) '', including '', including '', having '', including The terms "ins"), "containing", or any other variation thereof do not imply exclusive inclusion. It is intended to include. For example, a composition, mixture, or process consisting of a list of elements. The methods, articles, or apparatus are not necessarily limited to those elements alone. A composition, mixture, process, method, article, or apparatus that is explicitly listed as such It can include other elements that are not present.

[0065] Furthermore, the term “exemplary” in this specification means “functioning as an example, instance, or illustration.” It is used to mean "to do something". Any actual thing described as "exemplary" in this specification The implementation form or design is not necessarily preferable or advantageous to other embodiments or designs. It should not be interpreted as: The terms "at least one" and "one or more" should be interpreted as: The term "multiple" is understood to include any integer, such as 1, 2, 3, 4, etc. It is understood that this includes any integer greater than or equal to 2, that is, 2, 3, 4, 5, etc. The term can include both indirect and direct "connections."

[0066] The terms "approximately," "substantially," and "roughly," and their variations, may be used at the time of application. This is intended to include the degree of error associated with measuring a particular quantity based on capable instruments. For example, "approximately" can include a range of ±8%, 5%, or 2% of a specific value.

[0067] This invention relates to a system, method, or computer integrated at any possible level of technical detail. This can be a computer program product or a combination thereof. The product is a computer-readable program that causes the processor to execute an aspect of the present invention. It may include a computer-readable storage medium that stores the command.

[0068] Computer-readable storage media hold and store instructions used by instruction execution devices. It can be a tangible device capable of doing so. A computer-readable storage medium is one example. electronic memory devices, magnetic memory devices, optical memory devices, electromagnetic memory devices, semiconductor memory devices or Any suitable combination of these may be acceptable. A more specific example of a computer-readable storage medium. Examples include portable computer diskettes, hard disks, RAM, ROM, E PROM (or flash memory), SRAM, CD-ROM, DVD, memory stick A machine that records commands on a floppy disk, punch card, or grooved raised structure. Examples include mechanically encoded devices and appropriate combinations thereof. Computer-readable memory devices that use radio waves or other freely propagating electromagnetic waves, waveguides, etc. Electromagnetic waves that propagate through other transmission media (for example, light passing through an optical fiber cable) Transient signals themselves, such as pulses or electrical signals transmitted over a wire It should not be interpreted in that way.

[0069] The computer-readable program instructions described herein are not available from a computer-readable storage medium. It can be downloaded to each computer device / processor. Alternatively, it can be downloaded via the network. (For example, the internet, LAN, WAN or wireless network or this (These combinations) can be downloaded to an external computer or external storage device. The network consists of copper transmission cables, optical transmission fibers, wireless transmission, and routers. , firewall, switch, gateway computer or edge server or These combinations can be provided. Network within each computer device / processing unit A network adapter card or network interface allows computers to access the network. Upon receiving a computer-readable program instruction, the computer reads the program instruction and processes it in the respective computer. Transfer data to a computer-readable storage medium in a computer device / processing device.

[0070] Computer-readable program instructions for carrying out the operation of the present invention are assembler instructions, Instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, Firmware instructions, state setting data, configuration data for integrated circuits, or small talk Object-oriented programming languages ​​such as C++, and the C programming language One or more procedural programming languages, including or similar programming languages. Source code or object code written in any combination of grammar languages It can be either D or D. Computer-readable program instructions are standalone. As a software package, it can be installed entirely on the user's computer, or partially. It can run on the computer. Or, partially on the user's computer. Partially on a remote computer, or completely on a remote computer or It can be run on the server. In the latter case, the remote computer includes LANs and WANs. It may connect to the user's computer via any type of network, or external network To the computer (for example, via the internet using an internet service provider) They may be connected as follows. In some embodiments, for example, a programmable logic circuit, Field-programmable gate arrays (FPGAs), programmable logic arrays (PLGAs) An electronic circuit including A) is customized for the purpose of carrying out an aspect of the present invention. Therefore, by utilizing the state information of computer-readable program instructions, It can execute data-readable program instructions.

[0071] Embodiments of the present invention are described herein as methods, apparatus (systems) according to embodiments of the present invention. (m), and flowcharts or block diagrams or similar for computer program products It is explained by referring to both a flowchart or a block diagram or both. In each block, and in the flowchart or block diagram or both A combination of multiple blocks can be executed using computer-readable program instructions. ru.

[0072] The above computer-readable program instructions are for general-purpose computers used to produce machines. , provided to the processor of a dedicated computer or other programmable data processing device Good. This will enable the professional operation of such computers or other programmable data processing devices. These instructions, executed via a cessor, are used in flowcharts or block diagrams or similar formats. To create a means to perform a function / action identified in one or more blocks in both To output. The above computer-readable program instructions are further programmed by computers. Functions in a specific manner with respect to a data processing device or other device or a combination thereof. The instructions may be stored in a computer-readable storage medium that can be commanded to do so. The computer-readable storage medium that was used contains a flowchart or block diagram or Instructions for performing a mode of function / operation specified in one or more blocks in both. The product comprises products that include this.

[0073] Furthermore, computer-readable program instructions can be used by computers, other programmable devices, Alternatively, load it into another device and perform a series of operational steps on the computer, other programmer, etc. By running it on a device or another device, a computer execution process is generated. This may be done on the computer, other programmable device, or other device. The instructions executed are one or more in the flowchart, block diagram, or both. Execute the function / action specified in the block above.

[0074] The flowcharts and block diagrams in the drawings of this disclosure are for various embodiments of the present invention. Architectural representations of possible implementations of the system, method, and computer program product relating to the said It shows the functionality and operation. In this regard, flowcharts or blocks Each block in the diagram contains one or more executable instructions to perform a specific logical function. It can represent a module, segment, or part of an instruction, including other implementations. In terms of configuration, the functions shown within a block can be executed in a different order than the order shown in each figure. Good. For example, two blocks shown consecutively are actually related to the function they represent. They can be executed almost simultaneously, or in some cases, in reverse order. Note that a block diagram or... Each block in a flowchart or both, and a block diagram or flow - A combination of multiple blocks in a chart or both can have a specific function or movement. Dedicated hardware that performs operations or executes combinations of dedicated hardware and computer instructions. It is executable by a software-based system.

[0075] The descriptions of various embodiments of the present invention are presented for illustrative purposes only and are not intended to be exhaustive, nor are they intended to limit the disclosed embodiments. It will be apparent to those skilled in the art that many modifications and changes are possible without departing from the scope and spirit of the described embodiments. The terms used herein have been chosen to describe the principles of the embodiments, their practical application to market-based technologies or technical improvements, or to enable those skilled in the art to understand the embodiments herein. This disclosure also discloses the following inventions: <Note 1> The determination of an event related to a computing system, wherein the event occurs in a first time period, To acquire system data related to the aforementioned computing system, Based on the system data, the system state of the computing system at the first time is determined, Based on the system state, determine two or more system data clusters containing clustered system data related to the system state of the computing system, The interpolation algorithm determines the interpolated data value for the first time based on the system data, A computer implementation method comprising adjusting the interpolated data values ​​based on the determination that the interpolated data values ​​are outside of the two or more system data clusters. <Note 2> The computer implementation method according to Appendix 1, further comprising returning the interpolated data value based on the determination that the interpolated data value is located in at least one of the two or more system data clusters. <Note 3> Adjusting the interpolated data values ​​means The process involves determining the nearest system data cluster among the two or more system data clusters mentioned above, Determining a data value that lies midway between the interpolated data value and the centroid of the nearest system data cluster, wherein the data value is located within the nearest system data cluster. The computer implementation method according to Appendix 1, comprising generating an adjusted interpolated data value that includes the data value intermediate between the interpolated data value and the centroid of the nearest data cluster. <Note 4> Adjusting the interpolated data values ​​means The process involves determining the nearest system data cluster among the two or more system data clusters mentioned above, The process involves determining a data value that lies between the interpolated data value and the centroid of the nearest system data cluster, wherein the data value is located outside the nearest system data cluster. The computer implementation method according to Appendix 1, comprising generating an adjusted interpolated data value that includes the cluster boundary value of the nearest nearest data cluster, wherein the cluster boundary value is located within the nearest nearest data cluster and includes the data value closest to the interpolated data value. <Note 5> Determining the interpolated data values ​​for the first time based on the system data via an interpolation algorithm is: Determining a time segment that includes the period including the first time, Determining the system data values ​​within the aforementioned time segment, The computer implementation method according to Appendix 1, comprising performing interpolation to determine the interpolated data value using the system data value within the time segment via the interpolation algorithm. <Note 6> Determining the two or more system data clusters based on the aforementioned system state is: This involves analyzing past system data, including past system data values ​​and system state data, Using a clustering algorithm, the aforementioned past system data values ​​are placed into two or more system data clusters, The computer implementation method described in Appendix 1, comprising assigning a system state to each of the two or more system data clusters. <Note 7> The computer implementation method described in Appendix 1, wherein the interpolation algorithm includes a linear regression algorithm. <Note 8> Memory with computer-readable instructions, One or more processors for executing the computer-readable instruction, wherein the computer-readable instruction controls the one or more processors, The determination of an event related to a computing system, wherein the event occurs in a first time period, To acquire system data related to the aforementioned computing system, Based on the system data, the system state of the computing system at the first time is determined, Based on the system state, determine two or more system data clusters containing clustered system data related to the system state of the computing system, The interpolation algorithm determines the interpolated data value for the first time based on the system data, A system that performs an operation including adjusting the interpolated data value based on the fact that it has been determined that the interpolated data value is outside of the two or more system data clusters. <Note 9> The aforementioned operation is, The system according to Appendix 8, further comprising returning the interpolated data value based on the determination that the interpolated data value is located in at least one of the two or more system data clusters. <Note 10> Adjusting the interpolated data values ​​means The process involves determining the nearest system data cluster among the two or more system data clusters mentioned above, Determining a data value that lies midway between the interpolated data value and the centroid of the nearest system data cluster, wherein the data value is located within the nearest system data cluster. The system according to Appendix 8, comprising generating an adjusted interpolated data value that includes the data value intermediate between the interpolated data value and the centroid of the nearest data cluster. <Note 11> Adjusting the interpolated data values ​​means The process involves determining the nearest system data cluster among the two or more system data clusters mentioned above, The process involves determining a data value that lies between the interpolated data value and the centroid of the nearest system data cluster, wherein the data value is located outside the nearest system data cluster. The system according to Appendix 8, which includes generating an adjusted interpolation system data value that includes the cluster boundary value of the nearest nearest data cluster, wherein the cluster boundary value is located within the nearest nearest data cluster and includes the data value closest to the interpolation data value. <Note 12> Determining the interpolated data values ​​for the first time based on the system data via an interpolation algorithm is: Determining a time segment that includes the period including the first time, Determining the system data values ​​within the aforementioned time segment, The system according to Appendix 8, comprising performing interpolation to determine the interpolated data value using the system data value within the time segment via the interpolation algorithm. <Note 13> Determining the two or more system data clusters based on the aforementioned system state is: This involves analyzing past system data, including past system data values ​​and system state data, Using a clustering algorithm, the aforementioned past system data values ​​are placed into two or more system data clusters, The system as described in Appendix 8, which includes assigning a system state to each of the two or more system data clusters. <Note 14> The system described in Appendix 8, wherein the interpolation algorithm includes a linear regression algorithm. <Note 15> A computer program product including a computer-readable storage medium implementing program instructions, wherein the program instructions are executable by one or more processors, The determination of an event related to a computing system, wherein the event occurs in a first time period, To acquire system data related to the aforementioned computing system, Based on the system data, the system state of the computing system at the first time is determined, Based on the system state, determine two or more system data clusters containing clustered system data related to the system state of the computing system, The interpolation algorithm determines the interpolated data value for the first time based on the system data, A computer program product that causes the computer program to perform an operation including adjusting the interpolated data values ​​based on the fact that it has been determined that the interpolated data values ​​are outside of the two or more system data clusters. <Note 16> The aforementioned operation is, The computer program product according to Appendix 15, further comprising returning the interpolated data value based on the determination that the interpolated data value is located in at least one of the two or more system data clusters. <Note 17> Adjusting the interpolated data values ​​means The process involves determining the nearest system data cluster among the two or more system data clusters mentioned above, Determining a data value that lies midway between the interpolated data value and the centroid of the nearest system data cluster, wherein the data value is located within the nearest system data cluster. The computer program product according to Appendix 15, comprising generating an adjusted interpolation system data value that includes the data value intermediate between the interpolated data value and the centroid of the nearest data cluster. <Note 18> Adjusting the interpolated data values ​​means The process involves determining the nearest system data cluster among the two or more system data clusters mentioned above, The process involves determining a data value that lies between the interpolated data value and the centroid of the nearest system data cluster, wherein the data value is located outside the nearest system data cluster. The computer program product according to Appendix 15, which includes generating an adjusted interpolation system data value that includes the cluster boundary value of the nearest nearest data cluster, wherein the cluster boundary value is located within the nearest nearest data cluster and includes the data value closest to the interpolation data value. <Note 19> Determining the interpolated data value at the first time based on the system data via an interpolation algorithm comprises: Determining a time segment including a period that includes the first time; Determining system data values within the time segment; Executing an interpolation for determining the interpolated data value using the system data values within the time segment via the interpolation algorithm, the computer program product according to appendix 15. <Appendix 20> Determining the two or more system data clusters based on the system state comprises: Analyzing past system data including past system data values and system state data; Placing the past system data values into two or more system data clusters using a clustering algorithm; Assigning a system state to each of the two or more system data clusters, the computer program product according to appendix 15. <Appendix 21> Determining system data related to a computing system, the system data including a plurality of performance metric values related to the computing system acquired in a plurality of time slices; Generating a performance robustness value for each of the plurality of performance metric values using a neural network model; Plotting the performance robustness value for each of the plurality of performance metric values on an orthogonal plane; Clustering the performance robustness values into at least one system state of the computing system; Generating a data cluster for each of the at least one system state, the data cluster being associated with a performance metric of the computing system, the computer-implemented method. <Appendix 22> Receiving a set of performance metric values of the computing system; Interpolating a first performance metric value from the set of performance metric values; Determining a system state of the computing system at the first time; Adjusting the first performance metric value based on a determination that the first performance metric value is outside a first data cluster associated with the system state, the computer-implemented method according to Supplementary Note 21, further comprising. <Supplementary Note 23> Returning the first performance metric value based on a determination that the first performance metric value is within a first data cluster associated with the system state, the computer-implemented method according to Supplementary Note 22, further comprising. <Supplementary Note 24> A memory having computer-readable instructions; One or more processors for executing the computer-readable instructions, the computer-readable instructions controlling the one or more processors to Determining system data related to a computing system, the system data including a plurality of performance metric values related to the computing system acquired in a plurality of time slices; Using a neural network model to generate a performance robustness value for each of the plurality of performance metric values; Plotting the performance robustness value for each of the plurality of performance metric values on an orthogonal plane; Clustering the performance robustness values into at least one system state of the computing system; Generating a data cluster for each of the at least one system state, the data cluster being associated with a performance metric of the computing system, a system for executing an operation including generating. <Supplementary Note 25> The operation is Receiving a set of performance indicator values ​​for the aforementioned computing system, Interpolating a first performance index value from the set of performance index values, Determining the system state of the computing system at the first time, The system according to Appendix 24, further comprising adjusting the first performance metric value based on the determination that the first performance metric value is outside the first data cluster associated with the system state.

Claims

1. On the computer, Determining system data related to a computing system, wherein the system data includes multiple performance indicator values ​​related to the computing system acquired in multiple time slices. Using a neural network model, a performance robustness value is generated for each of the aforementioned multiple performance indicator values. For each of the aforementioned multiple performance indicator values, the performance robustness value is plotted on an orthogonal plane. The performance robustness value is clustered to at least one system state of the computing system, A method for generating and performing the following: generating a data cluster for each of the at least one system states, wherein the data cluster is associated with a performance metric of the computing system.

2. To the aforementioned computer, Receiving a set of performance indicator values ​​for the aforementioned computing system, Interpolating a first performance index value from the set of performance index values, Determining the system state of the computing system at the first time, The method according to claim 1, further comprising adjusting the first performance metric value based on the determination that the first performance metric value is outside the first data cluster associated with the system state.

3. To the aforementioned computer, The method of claim 2, further comprising: returning the first performance indicator value based on the determination that the first performance indicator value is in a first data cluster associated with the system state.

4. Memory with computer-readable instructions, One or more processors for executing the computer-readable instruction, wherein the computer-readable instruction controls the one or more processors, Determining system data related to a computing system, wherein the system data includes multiple performance indicator values ​​related to the computing system acquired in multiple time slices. Using a neural network model, a performance robustness value is generated for each of the aforementioned multiple performance indicator values. For each of the aforementioned multiple performance indicator values, the performance robustness value is plotted on an orthogonal plane. The performance robustness value is clustered to at least one system state of the computing system, A system that performs an operation including generating a data cluster for each of the at least one system states, wherein the data cluster is associated with a performance metric of the computing system.

5. The aforementioned operation is, Receiving a set of performance indicator values ​​for the aforementioned computing system, Interpolating a first performance index value from the set of performance index values, Determining the system state of the computing system at the first time, The system according to claim 4, further comprising adjusting the first performance metric value based on the determination that the first performance metric value is outside the first data cluster associated with the system state.

6. On the computer, Determining system data related to a computing system, wherein the system data includes multiple performance indicator values ​​related to the computing system acquired in multiple time slices. Using a neural network model, a performance robustness value is generated for each of the aforementioned multiple performance indicator values. For each of the aforementioned multiple performance indicator values, the performance robustness value is plotted on an orthogonal plane. The performance robustness value is clustered to at least one system state of the computing system, A computer program that generates and performs the following: generating data clusters for each of the at least one system states, wherein the data clusters are associated with performance metrics of the computing system.

Images