Resource elastic scaling method and system based on average consistency of multi-agent system

Through the average consistency control of multi-agent systems, dynamic event triggering and load observers are designed, which solves the high consumption of communication resource elasticity and network attacks in cloud computing, and achieves secure and distributed resource elasticity.

CN120523600APending Publication Date: 2025-08-22SHANDONG COMP SCI CENTNAT SUPERCOMP CENT IN JINAN +1
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202510661878.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-05-22
Publication Date
2025-08-22

AI Technical Summary

Technical Problem

The existing resource elastic scaling methods have problems such as high communication resource consumption, dependence on global information on communication topology, and inability to resist network attacks in cloud computing. Especially under distributed sequence scaling attacks, secure resource elastic scaling cannot be achieved.

Method used

Adopting the average consistency control based on multi-agent system, dynamic event triggering strategies based on edges and nodes are designed, combining load observers and time-varying feedback gains to achieve fully distributed, discrete communication and control, avoid unnecessary communication and resource consumption, and resist distributed sequence scaling attacks.

Benefits of technology

The server cluster's secure resource elastic scaling under distributed sequence scaling attack is realized, and the server load is stable on the average, reducing communication frequency, reducing resource consumption, avoiding misleading, and not relying on global information of communication topology.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120523600A_ABST
    Figure CN120523600A_ABST
Patent Text Reader

Abstract

The invention belongs to the technical field of computers, and discloses a resource elastic scaling method and system based on the average consistency of a multi-agent system, and the method comprises the steps: obtaining the state information of a current server, and estimating user input based on the state information and a scaling decision at the last moment; an event trigger is designed, and if the event trigger based on the edge is triggered, the server sends own state information to the neighbor and receives the state information of the neighbor; if the event trigger based on the node is triggered, the server records the state of the load observer of the server and the information of the self-adaptive variable; and forming a scaling decision according to the state information of the server at the last triggering moment, the state information of the neighbor server and the state information of the load observer, executing the scaling decision and updating the state information of the server. According to the method, on the premise that global information of communication topology is not used, the security resources of the server cluster under DSS attack are elastically expanded and contracted, and the server cannot be misguided by tampered information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a resource elastic scaling method and system based on average consistency of a multi-agent system. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In recent years, with the rapid development of computer technology and the explosive growth in data processing needs, enterprises and individual users have an increasing demand for computing power. However, traditional computing models suffer from inflexible resource allocation and low utilization, making them unable to effectively meet ever-changing workload demands. Against this backdrop, cloud computing, with its on-demand provisioning and flexible scalability, has become a crucial technology for meeting diverse computing needs. Resource elasticity is a core capability of cloud computing, designed to dynamically adjust the allocation of computing, storage, and other resources based on real-time workload changes to maximize resource utilization, reduce costs, and ensure service performance. Resource elasticity ensures that resource utilization on each server is maintained at an appropriate level: excessive resource utilization can lead to reduced work efficiency, while insufficient resource utilization can result in resource waste.

[0004] Compared to traditional control systems, multi-agent systems connect multiple subsystems via communication networks and achieve collaborative operation through information exchange. These systems offer significant advantages in terms of interoperability, fault tolerance, and scalability, and are widely used in fields such as cloud resource management. The consensus problem in multi-agent systems is a fundamental issue in the study of distributed cooperative control of multi-agent systems. Leaderless consensus typically requires that the states of all agents converge to the same value. The average consensus problem is particularly important in the study of leaderless consensus. The goal of this control problem is to design a control protocol for each agent to ensure that its state ultimately stabilizes at the average of the states of all agents. Furthermore, many existing control protocols rely on global information about the communication topology (such as the eigenvalues ​​of the Laplacian matrix or the number of agents). This means that if the communication topology of the multi-agent system network changes, these protocols may become ineffective. Furthermore, when the scale of a multi-agent system becomes large, global information about the communication topology is often difficult to obtain. Therefore, research on scalable, fully distributed control protocols has important theoretical and practical applications.

[0005] Existing resource scalers typically use deep learning techniques to learn time-series load information from server clusters, train models to predict future load trajectories, and use the predictions to make scaling decisions accordingly. Resource scaling strategies based on time series predictions achieve higher prediction accuracy and achieve better scaling decisions than traditional resource scaling strategies. However, existing methods typically incur high computational costs. If predictions deviate, resource scaling strategies based on time series predictions struggle to promptly correct scaling decisions.

[0006] In most cases, servers in a server cluster exchange information via network communication. Server autoscalers should minimize resource consumption when making scaling decisions, freeing up resources for executing the server's core tasks. However, many existing control strategies based on continuous communication require significant communication resources. For multi-agent systems, to avoid unnecessary communication and reduce communication resource consumption, node-based event-triggered strategies, following the traditional periodic sampling communication approach, have been introduced into the design of consensus control strategies. However, static event-triggered strategies with constant trigger thresholds have a drawback: they cannot guarantee asymptotic convergence of multi-agent consensus errors. Furthermore, in these node-based event-triggered strategies, each agent corresponds to an event trigger. At the moment of its event triggering, the agent communicates with all its neighboring agents, which also results in significant resource waste. In addition to inter-server communication, servers also consume resources when implementing scaling strategies to reallocate resources, and the amount of resources consumed can even be far greater than that consumed by inter-server communication. Therefore, ensuring that resource scalers make discrete scaling decisions to avoid unnecessary resource allocation is crucial for the practical application of resource scaling strategies.

[0007] In real-world scenarios, the communication environment between servers in a server cluster is less than ideal, and cyberattacks are common. These attacks can prevent the autoscaler from making timely and accurate scaling decisions. Therefore, elastically scaling resources in cloud-based server clusters under cyberattacks has become a pressing issue. Summary of the Invention

[0008] In order to solve the above problems, the present invention proposes a resource elastic scaling method and system based on the average consistency of a multi-agent system, which can realize the secure resource elastic scaling of a server cluster under DSS attacks without using the global information of the communication topology, and the server will not be misled by the tampered information.

[0009] In order to achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention provides a resource elastic scaling method based on average consistency of a multi-agent system, which is used to elastically control the computing power resources of a cloud-hosted server cluster, comprising the following steps: Obtain the current server status information and estimate user input based on the server status information and the scaling decision made at the previous moment. Design edge-based event triggers and node-based event triggers. If an edge-based event trigger is triggered, the server sends its own status information to neighboring servers and receives status information from neighboring servers. If a node-based event trigger is triggered, the server records its own load observer status and adaptive variable information at the current moment; The scaling decision at the current moment is formed based on the server's own state information at the last triggering moment, the state information of neighboring servers, and the state information of the load observer. The scaling decision at the current moment is executed and the server's own state information is updated.

[0010] As an optional implementation, an edge-based event trigger is specifically: An event trigger condition is set for each communication channel between servers in the server cluster. When the event is triggered, only the servers on both sides of the communication channel communicate and transmit their own status information to each other.

[0011] As an optional implementation, the event trigger based on the node is as follows: An event trigger condition is set for each server in the server cluster. When the event is triggered, the server records its load observer state and adaptive variable information at the current moment.

[0012] As an optional implementation manner, the thresholds of the event triggering conditions of the edge-based event trigger and the node-based event trigger are both dynamically changed.

[0013] As an optional implementation, consider the situation where the communication channel between servers in a server cluster hosted in the cloud is subject to a distributed sequence scaling attack. The distributed sequence scaling attack is modeled as a sequence form, and the frequency and duration of the distributed sequence scaling attack are required to meet certain conditions.

[0014] As an optional implementation, a time-varying feedback gain is set. When it is detected that the server cluster is under a distributed sequence scaling attack, the feedback gain is reset to zero, so that all servers affected by the distributed sequence scaling attack stop generating scaling policies.

[0015] In a second aspect, the present invention provides a resource elastic scaling system based on average consistency of a multi-agent system, comprising: Physical system, i.e. server system; A load observer estimates user input based on the server's own state information and the scaling decision made at the previous moment; Event triggers, including edge-based event triggers and node-based event triggers, are used to control the communication between adjacent servers and record the server's own information respectively; Zero-order holder, used to store information about the server itself and neighboring servers; The autoscaler is used to make scaling decisions at the current moment based on the server's own state information at the last trigger moment, the state information of neighboring servers, and the state information of the load observer.

[0016] In a third aspect, the present invention provides an electronic device comprising a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein the computer instructions, when executed by the processor, perform the method described in the first aspect.

[0017] In a fourth aspect, the present invention provides a computer-readable storage medium for storing computer instructions, wherein when the computer instructions are executed by a processor, the method described in the first aspect is performed.

[0018] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which implements the method described in the first aspect when executed by a processor.

[0019] Compared with the prior art, the present invention has the following beneficial effects: This paper proposes a resource elastic scaling method and system based on average consistency of a multi-agent system. This method and system address distributed sequence scaling (DSS) attacks against a server cluster with an undirected communication topology. Through secure average consistency control of the multi-agent system, this method achieves secure resource elastic scaling of the server cluster using a fully distributed, discrete communication, and discrete control approach. This method ensures that the load of each server in the server cluster ultimately stabilizes at the average load of all servers. This method achieves secure resource elastic scaling of the server cluster under DSS attacks without using global information about the communication topology, and servers are protected from being misled by tampered information.

[0020] The present disclosure proposes a resource elastic scaling method and system based on the average consistency of a multi-agent system. By introducing the frequency and duration of attacks into the attack model, a DSS attack model is constructed. A load observer is designed to estimate external interference, and the state of the load observer is introduced into the scaling strategy to offset the impact of external interference. An edge-based dynamic event triggering strategy is designed for communication between servers, which avoids unnecessary communication between servers and reduces the communication frequency. While excluding Zeno behavior, it ensures that the strategy can achieve safe average consistency of the server cluster in a discrete communication manner. A node-based dynamic event triggering strategy is designed for the autoscaler, which reduces resource consumption and achieves discrete updates of scaling decisions while excluding Zeno behavior. The proposed safe resource elastic scaling strategy does not rely on global information of the communication topology and can operate in a fully distributed, discrete communication and discrete control manner.

[0021] Advantages of additional aspects of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0023] Figure 1 Schematic diagram of the structure of the resource elastic scaling system based on the average consistency of the multi-agent system of the present invention; Figure 2 This is a flow chart of the resource elastic scaling method based on average consistency of a multi-agent system according to the present invention; Figure 3 A trajectory diagram of the CPU utilization of each server of the present invention; Figure 4 A trajectory diagram of the response time of each server of the present invention; Figure 5 The event triggering time of each communication channel in the server cluster of the present invention; Figure 6 This is the scaling decision for some servers in the server cluster of the present invention. DETAILED DESCRIPTION

[0024] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0025] It should be noted that the following detailed descriptions are exemplary and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present invention belongs.

[0026] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0027] In the absence of conflict, the embodiments of the present invention and the features thereof may be combined with each other.

[0028] Explanation of terms: 1. Multi-agent systems: A multi-agent system is composed of multiple agents, each of which can independently perceive its environment and make decisions, and can communicate and collaborate with other agents to achieve a goal. Agents are entities that can perceive their surroundings and act according to certain rules or strategies. They typically possess capabilities such as perception, reasoning, learning, and decision-making.

[0029] 2. Neighbors / adjacent agents: Agents that are connected to agent i in the communication topology, that is, agents that can communicate with agent i, are called neighbors / adjacent agents of agent i.

[0030] 3. Event-triggered strategy: In an event-triggered strategy, whether agents communicate with each other is determined by a pre-designed event trigger condition consisting of the error in their current state. If the trigger condition is violated at a certain moment, the agents communicate; otherwise, they do not. A node-based event-triggered strategy means that each agent corresponds to an event trigger condition. When an event is triggered, the agent communicates with all neighboring agents. In contrast, an edge-based event-triggered strategy means that each communication channel corresponds to an event trigger condition. When an event is triggered, only the agents on both sides of the communication channel communicate.

[0031] 4. Distributed Sequence Scaling Attacks (DSS): A distributed sequence scaling attack is a type of distributed spoofing attack. A spoofing attack on a communication channel modifies the data transmitted between agents on either side of the channel, rendering it unreliable. A scaling attack modifies the data transmitted between agents into the product of the original data and a scaling factor. A distributed sequence scaling attack is a scaling attack that is launched in a time series (i.e., it persists for a period of time after initiation), allowing it to independently attack any communication channel at any time.

[0032] 5. Resource Scaling: Resource scaling refers to the system dynamically adjusting the allocation of compute, storage, network, and other resources based on current load conditions to improve resource utilization and ensure system performance. Scaling mechanisms typically determine when to expand (add resources) or contract (release resources) based on preset policies or real-time monitoring data. Scaling triggers can include key metrics such as CPU utilization, memory usage, and request response time. Node-based scaling means each compute node independently decides whether to scale, while cluster-based scaling adjusts resource allocation across the entire system to meet changing business needs.

[0033] 6. Zeno behavior: A phenomenon in which an event is triggered an infinite number of times within a finite period of time.

[0034] Example 1 like Figure 2 As shown, this embodiment provides a resource elastic scaling method based on the average consistency of a multi-agent system, which is used to perform elastic scaling control of computing power resources on a server cluster hosted in the cloud, including the following steps: Obtain the current server status information and estimate user input based on the server status information and the scaling decision made at the previous moment. Design edge-based event triggers and node-based event triggers. If an edge-based event trigger is triggered, the server sends its own status information to neighboring servers and receives status information from neighboring servers. If a node-based event trigger is triggered, the server records its own load observer status and adaptive variable information at the current moment; The scaling decision at the current moment is formed based on the server's own state information at the last triggering moment, the state information of neighboring servers, and the state information of the load observer. The scaling decision at the current moment is executed and the server's own state information is updated.

[0035] The invention models the cloud-hosted server cluster as a multi-agent system, models user input as external interference, and models the resource elastic scaling problem of the server cluster as the average consistency problem of the multi-agent system subject to external interference, thereby achieving elastic scaling of the server cluster.

[0036] The state of the agent is used to represent the resource utilization status of each server in the server cluster (such as CPU utilization , response time wait), Figure 3 The CPU utilization trajectory of each server is shown in the figure. Figure 4 The response time trajectory of each server is plotted. The control input is used to represent the scaling decision made for each server (such as the number of virtual machines). etc.), using external interference to represent user input to each server The multi-agent system modeled according to the example is as follows:

[0037] We use N agents to represent N servers in a server cluster and describe the dynamics of N agents in a multi-agent system as:

[0038] in Representing an agent The state of the server resource utilization; Representing an agent The control input of the server Scaling decisions; Representing an agent External interference, that is, the server User input. and is the system matrix, and Is the input matrix of external disturbance. External disturbance Produced by the following system:

[0039] in, is the system matrix of the external system, for The derivative function of .

[0040] Consider the case where the communication channel between servers in a server cluster hosted in the cloud is attacked by DSS. Model the DSS attack as a sequence, where the communication channel On the DSS attacks in Started at time, duration is Therefore, let Indicates Occurs within the communication channel The set of time intervals of DSS attacks on Indicates internal communication channel The set of time intervals that have not been attacked by DSS. The present invention adopts an event triggering strategy, under which adjacent agents will only communicate and transmit information at the time of event triggering. This means that if the data transmitted between agents is tampered with by the DSS attack at the time of the previous event triggering, the agents will not be able to transmit the original data that has not been tampered with until the next event triggering time that has not been attacked by DSS. Therefore, the time interval actually affected by the DSS attack is greater than the time interval during which the DSS attack lasts. Let Indicates the communication channel Based on this, we will exist The set of time intervals actually affected by the DSS attack is expressed as follows:

[0041] Accordingly, the set of time intervals not affected by the DSS attack is expressed as follows:

[0042] make Indicates the communication channel exist Frequency of DSS attacks in Indicates the communication channel exist The number of DSS attacks in . Frequency of internal DSS attacks satisfy:

[0043] in, Indicates the communication channel at the initial moment The number of DSS attacks.

[0044] for internal communication channel Total duration of DSS attacks ,satisfy:

[0045] in, Indicates Moment, the duration of the DSS attack; Indicates the ratio of the total duration of a DSS attack to the total duration.

[0046] To achieve fully distributed secure resource elastic scaling of cloud-hosted server clusters under DSS attacks, the present invention designs the following secure resource elastic scaler:

[0047] in represents the feedback matrix to be determined, The elements at the corresponding positions in the adjacency matrix representing the system communication topology. The remaining items will be introduced in detail in the subsequent content.

[0048] The present invention models user input as external disturbance and estimates the external disturbance through a load observer. The specific form of the load observer is as follows:

[0049] in, is the observer state, which means that for the agent An estimate of the external disturbance; is the observer gain to be determined, for The derivative function of .

[0050] For inter-server communication, the present invention adopts an edge-based event triggering strategy. Each communication channel corresponds to an event triggering condition. When an event is triggered, only the agents on both sides of the communication channel communicate. Furthermore, the thresholds in the event triggering conditions are dynamically variable. Figure 5 The event triggering time for each communication channel in the server cluster.

[0051] We order , Indicates on the edge On the Event triggering moment. For undirected communication channel ,server When the event is triggered Actively send to the server Send your current status information and read the server State information at the current moment. Considering the impact of DSS attacks on the information transmitted between servers, we assume For the edge The scaling factor of the DSS attack. In the interval between two adjacent event triggering, that is, When the present invention uses a zero-order holder to store the state information of the neighbor server at the time of the last event triggering, the specific form of the state information is as follows:

[0052] in, .

[0053] We use the following event trigger conditions to determine the triggering time of the above events , .

[0054]

[0055] in is a dynamic variable. and is a positive constant and . Event trigger function As shown below:

[0056] in , , is a constant matrix to be determined, is a positive constant, is an adaptive variable that is dynamically as follows:

[0057] in, is a positive constant, is the scaling gain to be determined.

[0058] To achieve discrete updates to scaling policies, we designed a node-based static event triggering strategy for the resource scaler. The thresholds in the event triggering conditions also change dynamically.

[0059] We order , Representation Server No. The event triggers at that moment. Server Will be Read the state of the disturbance observer at all times and adaptive variables , and the zero-order holder also stores the above information. We define its specific form as:

[0060] Event triggering time , Determined by the following event trigger conditions:

[0061] Among them, the event trigger function As shown below:

[0062] in .

[0063] For the above security resource elastic scaling strategy, let 、 and , if there exists a positive constant 、 and , a positive definite matrix 、 and The following conditions must be met:

[0064] And there is a constant 、 and ,matrix and The frequency and duration of DSS attacks must meet the following conditions:

[0065] This automatic scaling strategy can achieve elastic scaling of security resources in the server cluster, and the event triggering strategy will not trigger Zeno behavior.

[0066] During a DSS attack, the tampered false information passed between agents may cause the controller to generate incorrect control inputs, thereby misleading the agents. Therefore, to prevent the tampered information from misleading the agents, we modify the autoscaler in the above autoscaling strategy to:

[0067] in, It is a time-varying feedback gain. When the system detects that it is under DSS attack, the feedback gain is set to zero, so that all servers affected by the DSS attack stop generating scaling strategies and stop adaptive variables. The update of tampered false information prevents the server from misleading. The specific form of the time-varying feedback gain is as follows:

[0068] For the above security resource elastic scaling strategy, let 、 and , if there exists a positive constant 、 and , a positive definite matrix 、 and The following conditions must be met:

[0069] And there is a constant 、 and ,matrix The frequency and duration of DSS attacks must meet the following conditions:

[0070] This automatic scaling strategy can achieve elastic scaling of security resources in the server cluster, and the event triggering strategy will not trigger Zeno behavior. Figure 6 Make scaling decisions for some servers in a server cluster.

[0071] The resource elastic scaling method based on the average consistency of a multi-agent system of the present invention is aimed at a server cluster with an undirected communication topology under distributed sequence scaling attacks (DSS). Through the secure average consistency control of the multi-agent system, the secure resource elastic scaling of the server cluster is achieved in a fully distributed, discrete communication and discrete control manner, so that the load of each server in the server cluster is finally stabilized at the average value of the load of all servers.

[0072] This secure autoscaling method takes DSS attacks into consideration and constructs a DSS attack model by introducing the frequency and duration of attacks into the attack model. To cope with external interference (i.e., user input), we design a load observer to estimate external interference and introduce the state of the load observer into the scaling strategy to offset the impact of external interference. To avoid unnecessary communication between servers and further reduce the communication frequency, we design an edge-based dynamic event trigger for inter-server communication. While excluding Zeno behavior, we ensure that this method can achieve secure average consistency of the server cluster in a discrete communication manner. To further reduce resource consumption, we design a node-based dynamic event trigger for the autoscaler. While excluding Zeno behavior, we achieve discrete updates of scaling decisions. The proposed secure resource elastic scaling method does not rely on global information about the communication topology and can operate in a fully distributed, discrete communication, and discrete control manner.

[0073] Example 2 like Figure 1 As shown, this embodiment provides a resource elastic scaling system based on the average consistency of a multi-agent system, including: Physical system, i.e. server system; A load observer estimates user input based on the server's own state information and the scaling decision made at the previous moment; Event triggers, including edge-based event triggers and node-based event triggers, are used to control the communication between adjacent servers and record the server's own information respectively; Zero-order holder, used to store information about the server itself and neighboring servers; The autoscaler is used to make scaling decisions at the current moment based on the server's own state information at the last trigger moment, the state information of neighboring servers, and the state information of the load observer.

[0074] As shown in Figure 1, the resource elastic scaling system is designed with six units within the agent (server): the physical system, the load observer, the edge-based event trigger, the node-based event trigger, the zero-order retainer, and the autoscaler. The agents are connected via a communication network.

[0075] It should be noted that the above modules correspond to the steps described in Example 1, and the examples and application scenarios implemented by the above modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above Example 1. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0076] In further embodiments, there is also provided: An electronic device includes a memory and a processor, and computer instructions stored in the memory and executed by the processor, wherein when the computer instructions are executed by the processor, the method described in Example 1 is performed. For the sake of brevity, no further details are given here.

[0077] It should be understood that in this embodiment, the processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), off-the-shelf field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0078] The memory may include a read-only memory and a random access memory, and provides instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0079] A computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor, the method described in Example 1 is performed.

[0080] The method in Example 1 can be directly implemented as a hardware processor, or can be implemented using a combination of hardware and software modules in the processor. The software module can be located in a storage medium mature in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, it will not be described in detail here.

[0081] A computer program product includes a computer program, which implements the method described in embodiment 1 when executed by a processor.

[0082] The present invention also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer-executable instructions, such as instructions contained in program modules, which are executed in a device on a real or virtual processor of a target to perform the process / method described above. Generally, program modules include routines, programs, libraries, objects, classes, components, data structures, etc. that perform specific tasks or implement specific abstract data types. In various embodiments, the functionality of program modules can be combined or divided between program modules as needed. The machine-executable instructions for the program modules can be executed in local or distributed devices. In distributed devices, program modules can be located in local and remote storage media.

[0083] The computer program code for implementing the method of the present invention can be written in one or more programming languages. These computer program codes can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the computer or other programmable data processing device, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on a computer, partially on a computer, as an independent software package, partially on a computer and partially on a remote computer, or entirely on a remote computer or server.

[0084] In the context of the present invention, computer program code or related data can be carried by any appropriate carrier to enable a device, apparatus, or processor to perform the various processes and operations described above. Examples of carriers include signals, computer-readable media, and the like. Examples of signals include electrical, optical, radio, acoustic, or other forms of propagated signals, such as carrier waves, infrared signals, and the like.

[0085] Those skilled in the art will appreciate that the units and algorithm steps of the various examples described in conjunction with this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0086] Although the above describes the specific embodiments of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without any creative work are still within the scope of protection of the present invention.

Claims

1. A resource elastic scaling method based on the average consistency of a multi-agent system, characterized by: Used to elastically scale computing resources for cloud-hosted server clusters, including the following steps: Obtain the current server status information and estimate user input based on the server status information and the scaling decision made at the previous moment. Design edge-based event triggers and node-based event triggers. If an edge-based event trigger is triggered, the server sends its own status information to neighboring servers and receives status information from neighboring servers. If a node-based event trigger is triggered, the server records its own load observer status and adaptive variable information at the current moment; The scaling decision at the current moment is formed based on the server's own state information at the last triggering moment, the state information of neighboring servers, and the state information of the load observer. The scaling decision at the current moment is executed and the server's own state information is updated.

2. The resource elastic scaling method based on average consistency of a multi-agent system according to claim 1, characterized in that: Edge-based event triggers, specifically: An event trigger condition is set for each communication channel between servers in the server cluster. When the event is triggered, only the servers on both sides of the communication channel communicate and transmit their own status information to each other.

3. The resource elastic scaling method based on average consistency of a multi-agent system according to claim 1, characterized in that: Node-based event triggers, specifically: An event trigger condition is set for each server in the server cluster. When the event is triggered, the server records its load observer state and adaptive variable information at the current moment.

4. The resource elastic scaling method based on average consistency of a multi-agent system according to claim 1, characterized in that: The thresholds of event triggering conditions for edge-based event triggers and node-based event triggers are both dynamically changing.

5. The resource elastic scaling method based on average consistency of a multi-agent system according to claim 1, characterized in that: Considering the situation where the communication channel between servers in a cloud-hosted server cluster is subjected to a distributed sequence scaling attack, the distributed sequence scaling attack is modeled as a sequence, and the frequency and duration of the distributed sequence scaling attack are required to meet certain conditions.

6. The resource elastic scaling method based on average consistency of a multi-agent system according to claim 1, characterized in that: A time-varying feedback gain is set. When it is detected that the server cluster is under a distributed sequence scaling attack, the feedback gain is reset to zero, so that all servers affected by the distributed sequence scaling attack stop generating scaling strategies.

7. Resource elastic scaling system based on average consistency of multi-agent system, characterized by: include: Physical system, i.e. server system; A load observer estimates user input based on the server's own state information and the scaling decision made at the previous moment; Event triggers, including edge-based event triggers and node-based event triggers, are used to control the communication between adjacent servers and record the server's own information respectively; Zero-order holder, used to store information about the server itself and neighboring servers; The autoscaler is used to make scaling decisions at the current moment based on the server's own state information at the last trigger moment, the state information of neighboring servers, and the state information of the load observer.

8. An electronic device, characterized in that: The method comprises a memory and a processor, and computer instructions stored in the memory and executed on the processor, wherein when the computer instructions are executed by the processor, the method according to any one of claims 1 to 6 is completed.

9. A computer-readable storage medium, characterized in that Used to store computer instructions, which, when executed by a processor, complete the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The invention comprises a computer program, which is used to implement the method according to any one of claims 1 to 6 when the computer program is executed by a processor.

Citation Information

Cited By

  • Method for realizing bilateral consistency of multi-agent system under sequence scaling attack

    CN121501018A

  • CDN (Content Delivery Network) edge node collaborative optimization method, system, equipment and medium

    CN122248033A

  • A CDN edge node collaborative optimization method, system, device and medium

    CN122248033B