Distributed cloud server load balancing optimization system and method based on edge computing

Abstract

The invention relates to the technical field of cloud servers, in particular to a distributed cloud server load balancing optimization system and method based on edge computing. The method comprises the following steps: scheduling a request signal sent by a user terminal to a preferred server with the shortest distance value from the user terminal for the first time, and obtaining a first utilization rate improvement degree of the whole cloud server after the first scheduling; secondarily scheduling the request signal to the alternative server, and obtaining the suitability of the secondary scheduling to the alternative server and the second utilization rate improvement degree of the secondary scheduling to the whole cloud server; the scheduling matching degree is obtained based on the suitability degree, the first utilization rate improvement degree and the second utilization rate improvement degree, if the maximum scheduling matching degree is a non-negative number, the alternative server corresponding to the maximum scheduling matching degree is the final server, otherwise, the preferred server is the final server, and the final server serves the request signal. According to the invention, the load scheduling strategy is more reasonable, and the resource utilization rate of the cloud server is higher.

Description

Technical Field

[0001] This invention relates to the field of cloud server technology, and more specifically to a distributed cloud server load balancing optimization system and method based on edge computing. Background Technology

[0002] Cloud servers leverage virtualization technology to pool large-scale physical resources and encapsulate them into elastically scalable computing units, providing efficient, secure, and easily managed computing services. This reduces users' operational complexity and overall IT costs, allowing them to focus more on core business. However, when user requests surge, server load increases, potentially leading to resource overload and service failure. Therefore, load balancing technology is needed to distribute tasks across multiple operating units to improve resource utilization and system throughput.

[0003] Traditional cloud servers are based on a centralized data center model, which distributes user requests to multiple servers within the data center to achieve load balancing and improve resource utilization and system throughput. However, when dealing with a sudden surge in user requests (such as high-speed rail ticketing or ticket grabbing), all data needs to be transmitted to the central server, resulting in high network latency, slow cloud server response, and enormous pressure on the central network bandwidth. This makes it difficult to meet real-time requirements and may even lead to server crashes.

[0004] To address the issues of high latency and potential failure of traditional centralized load balancing in cloud servers during sudden surges in user requests, existing technologies often employ edge computing for load balancing. This involves deploying edge servers in the peripheral areas surrounding the data source or users, routing user requests to the nearest edge server node to reduce network transmission distance, lower latency, and alleviate pressure on the cloud server center. However, blindly routing requests based solely on distance can easily lead to localized high loads on some edge servers while others remain idle. Summary of the Invention

[0005] To address the technical problem of low cloud server resource utilization due to inadequate current load scheduling strategies, this invention aims to provide a distributed cloud server load balancing optimization method based on edge computing. The specific technical solution adopted is as follows: Receive request signals sent by user terminals; The request signal is initially scheduled to the preferred server and the first utilization improvement of the overall cloud server after the initial scheduling is obtained. The preferred server is the edge server with the shortest distance to the user terminal. The request signal is rescheduled to a candidate server, and the suitability of the rescheduling to the candidate server and the second utilization improvement of the rescheduling to the overall cloud server are obtained. The candidate server is any edge server other than the preferred server. Based on the suitability, the first utilization improvement, and the second utilization improvement, a scheduling matching degree is obtained. If the scheduling matching degree with the largest value is a non-negative number, then the candidate server corresponding to it is the final server; otherwise, the preferred server is the final server, and the final server serves the request signal.

[0006] Furthermore, the process of obtaining the utilization improvement includes: The load dispersion of each preset hardware in the edge server is obtained, and the optimizability of the overall cloud server load scheduling is obtained based on the load dispersion. The utilization improvement is calculated by obtaining the load scheduling optimizability and the utilization rate of each preset hardware before and after the initial scheduling.

[0007] Furthermore, the process of obtaining the load dispersion includes: Calculate the absolute value of the difference between the utilization rate of the target hardware and the average utilization rate, wherein the target hardware is any of the preset hardware, and the average utilization rate is the average value of the utilization rate of the target hardware in the edge server; Calculate the coefficient of variation of the utilization rate of the target hardware; The load dispersion is calculated based on the ratio between the absolute value and the utilization rate of the corresponding target hardware, and the coefficient of variation.

[0008] Furthermore, the process of obtaining the optimizability of the load scheduling includes: The load scheduling optimizability is obtained by averaging the load dispersion of each preset hardware and then normalizing it.

[0009] Furthermore, the process of obtaining the utilization improvement degree also includes: The ratio of the load scheduling optimizability before the initial scheduling to the load scheduling optimizability after the initial scheduling is taken as the optimizability ratio. Based on the utilization rates of each preset hardware before and after the initial scheduling, the change in utilization rate of each preset hardware is obtained, and the average value of the change in utilization rate is calculated. The utilization improvement is obtained based on the average of the optimizability ratio and the utilization rate change magnitude.

[0010] Furthermore, the process of obtaining the utilization rate change range includes: For any hardware, the utilization rate after the initial scheduling is subtracted from the utilization rate before the initial scheduling to obtain the increase in utilization rate, and the increase in utilization rate is divided by the utilization rate after the initial scheduling to obtain the change in utilization rate for any hardware.

[0011] Furthermore, the process of obtaining the suitability includes: The ratio between the CPU utilization of the preferred server and the CPU utilization of the alternative servers is denoted as the CPU utilization ratio. Obtain the first distance between the preferred server and the user terminal, and the second distance between the alternative server and the user terminal; The suitability is obtained based on the CPU utilization ratio and the difference between the second distance and the first distance.

[0012] Furthermore, the preset hardware includes: CPU, memory, and I / O interface.

[0013] Furthermore, the method also includes: before receiving the request signal sent by the user terminal, initializing the cloud server configuration according to the user's needs, deploying a base station to form a wireless access network, connecting the user terminal through a communication network device, and deploying the edge server at the network edge to build an edge computing system.

[0014] This invention also provides a distributed cloud server load balancing optimization system based on edge computing, the system comprising: The receiving module is used to receive request signals sent by the user terminal; The initial scheduling module is used to initially schedule the request signal to the preferred server and obtain the first utilization improvement of the overall cloud server after the initial scheduling. The preferred server is the edge server with the shortest distance to the user terminal. The secondary scheduling module is used to reschedule the request signal to a candidate server and obtain the suitability of rescheduling to the candidate server and the second utilization improvement of the secondary scheduling on the overall cloud server. The candidate server is any edge server other than the preferred server. The final scheduling determination module is used to obtain a scheduling match degree based on the suitability, the first utilization improvement degree, and the second utilization improvement degree. If the scheduling match degree with the largest value is a non-negative number, then the candidate server corresponding to it is the final server; otherwise, the preferred server is the final server, and the final server serves the request signal.

[0015] The present invention has the following beneficial effects: First, a request signal sent by the user terminal is received. This request signal forms the basis for subsequent analysis.

[0016] Secondly, the request signal is initially scheduled to the preferred server, and the initial utilization improvement of the overall cloud server after the initial scheduling is obtained. The preferred server is the edge server with the shortest distance to the user terminal. The higher the initial utilization improvement, the higher the overall utilization improvement of the cloud server after the initial scheduling.

[0017] Furthermore, the request signal is rescheduled to a candidate server, and the suitability of this rescheduling and the second utilization improvement of the secondary scheduling on the overall cloud server are obtained. The candidate server is any edge server other than the primary server. A higher second utilization improvement indicates a higher overall utilization improvement of the cloud server due to the secondary scheduling.

[0018] Finally, a scheduling match score is obtained based on the suitability, the first utilization improvement score, and the second utilization improvement score. If the scheduling match score with the highest value is non-negative, the corresponding candidate server is the final server; otherwise, the preferred server is the final server, and the final server serves the request signal. When the scheduling match score with the highest value is non-negative, it indicates that the secondary scheduling has a greater improvement in the overall utilization of the cloud server compared to the initial scheduling. The higher the non-negative value of the scheduling match score with the highest value, the better the effect of selecting the candidate server corresponding to the scheduling match score with the highest value to serve the request signal.

[0019] In summary, by using this invention, the matching degree between user requests and each edge server can be comprehensively considered, load scheduling can be carried out reasonably, and the overall resource utilization of cloud servers can be improved. Attached Figure Description

[0020] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 A flowchart illustrating the distributed cloud server load balancing optimization method based on edge computing provided in the first embodiment of the present invention; Figure 2 A flowchart illustrating the process of obtaining the utilization improvement rate provided in the second embodiment of the present invention; Figure 3 A flowchart illustrating the process of obtaining load dispersion provided in the third embodiment of the present invention; Figure 4A flowchart illustrating the suitability acquisition process provided in the fourth embodiment of the present invention; Figure 5 This is a schematic diagram of a distributed cloud server load balancing optimization system based on edge computing provided in the fifth embodiment of the present invention. Figure 6 This is an example diagram of a distributed cloud server network architecture based on edge computing provided in the first embodiment of the present invention. Detailed Implementation

[0022] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the edge computing-based distributed cloud server load balancing optimization system and method proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] The main objective of this invention is: By comprehensively considering the matching degree between user requests and each edge server, load scheduling is carried out reasonably to improve the overall resource utilization of cloud servers.

[0025] The scenario addressed by this invention is: When scheduling edge servers to handle user request load, there may be a surge in the number of users in a local area. Scheduling edge servers solely based on physical distance can easily lead to situations where some edge servers are overloaded while others are idle, resulting in insufficient utilization of cloud server resources as a whole.

[0026] The specific scheme of the load balancing optimization method for distributed cloud servers based on edge computing provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0027] Please see Figure 1 The diagram illustrates a flowchart of a distributed cloud server load balancing optimization method based on edge computing provided in the first embodiment of the present invention. The method includes: S101. Receive the request signal sent by the user terminal.

[0028] The request signal carries the geographical location information of the user terminal.

[0029] Figure 6This is an example diagram of the distributed cloud server network architecture based on edge computing provided in the first embodiment of the present invention. Please refer to... Figure 6 In this invention, before receiving a request signal from a user terminal, the cloud server configuration is initialized according to user needs, and a base station is deployed to form a wireless access network. The user terminal is connected through communication network equipment, and the edge server is deployed at the network edge to construct an edge computing system. After deployment, the system has n edge servers, and the physical location information of each edge server in the network topology is recorded.

[0030] S102. The request signal is initially scheduled to the preferred server and the first utilization improvement of the overall cloud server after the initial scheduling is obtained. The preferred server is the edge server with the shortest distance to the user terminal.

[0031] The preferred server is the edge server with the shortest distance to the user terminal, where the shortest distance refers to the shortest straight-line distance.

[0032] The process of obtaining the utilization improvement will be described in detail in the second embodiment, and will not be repeated here.

[0033] S103. The request signal is rescheduled to a candidate server and the suitability of the rescheduled reschedule to the candidate server and the second utilization improvement of the rescheduled reschedule to the overall cloud server are obtained. The candidate server is any edge server other than the preferred server.

[0034] The process of obtaining the suitability will be described in detail in the fourth embodiment, and will not be repeated here.

[0035] The process of obtaining the second utilization improvement is the same as that of obtaining the first utilization improvement, and will not be described again here.

[0036] S104. Based on the suitability, the first utilization improvement degree and the second utilization improvement degree, obtain the scheduling matching degree. If the scheduling matching degree with the largest value is a non-negative number, then the candidate server corresponding to it is the final server. Otherwise, the preferred server is the final server, and the final server serves the request signal.

[0037] The scheduling matching degree can be expressed by the formula: ; Among them, the This indicates that the request signal is rescheduled to the alternative server, i.e., the first... The second utilization improvement of the edge server, the This indicates that the request signal is initially dispatched to the preferred server, i.e., the first... The first utilization improvement of the edge server, the This indicates that the request signal has been rescheduled to the second time. The suitability of each edge server, the This represents the maximum suitability of the request signal being secondary-scheduled to an edge server other than the preferred server. The subscript M indicates the meaning of the maximum value. This indicates that the request signal is related to the first Scheduling matching degree between edge servers.

[0038] The The value represents the first The relative suitability of the first edge server among all edge servers; the larger the value, the higher the suitability. Among all edge servers, the edge server is relatively more suitable for secondary scheduling of user requests.

[0039] The This indicates the improvement in the overall utilization of the cloud server caused by the secondary scheduling compared to the primary scheduling. The larger the difference, the greater the improvement of the secondary scheduling compared to the primary scheduling, and the higher the matching degree between the secondary scheduling and the current user request.

[0040] In summary, this user request is related to the first... The scheduling matching degree among the edge servers is The larger this value, the higher the matching degree between the two scheduling, and the more necessary it is to reschedule user requests to the edge server.

[0041] Due to the first The edge server selected by the secondary scheduling can be any edge server other than the preferred server. Given a total of n edge servers, therefore, the... The value can have n-1 possible values, the aforementioned The maximum value can be denoted as .

[0042] like If the corresponding edge server is considered to have a high degree of matching with the current user request and a higher degree of improvement in the overall utilization of the cloud server, the current user request will be rescheduled to the edge server for service, thus completing the load scheduling task.

[0043] In other words, when the scheduling match degree is the largest and is non-negative, the corresponding alternative server (the edge server selected in the secondary scheduling) is the final server; when the scheduling match degree is the largest and is negative, the preferred server (the edge server selected in the initial scheduling) is the final server, and the final server serves the request signal.

[0044] Figure 2 The flowchart illustrates the process of obtaining the utilization improvement rate according to the second embodiment of the present invention. The process of obtaining the utilization improvement rate includes: S201. Obtain the load dispersion of each preset hardware in the edge server, and obtain the optimizability of the overall cloud server load scheduling based on the load dispersion.

[0045] The server processes user requests simultaneously using a parallel operation. After a user request is initially scheduled to the corresponding server, the load on the server's internal hardware increases due to processing the request (e.g., CPU utilization, memory usage, and I / O throughput). Therefore, the load requirement of the user request can be determined by observing the changes in the server's hardware utilization before and after scheduling.

[0046] Among them, CPU utilization can be the average utilization over one minute, memory usage can be the average usage over one minute, and IO throughput can be the average IO throughput over one minute.

[0047] The process of obtaining the load dispersion will be described in detail in the third embodiment, and will not be repeated here.

[0048] The process of obtaining the optimizability of the load scheduling includes: The load scheduling optimizability is obtained by averaging the load dispersion of each preset hardware and then normalizing it.

[0049] The preset hardware may include: CPU, memory, I / O interfaces, etc., totaling... For each type of hardware, the load dispersion is obtained using the method described above and denoted as follows: The higher the load dispersion of each hardware component, the greater the optimization space for load scheduling. Let the normalized average value of all load dispersions be denoted as... The As a measure of the overall optimizability of cloud server load scheduling, the aforementioned The average value representing the dispersion of all loads is... This represents a normalization function, such as the range normalization function.

[0050] S202. Calculate the utilization improvement by obtaining the load scheduling optimizability and the utilization rate of each preset hardware before and after the initial scheduling.

[0051] The ratio of the load scheduling optimizability before the initial scheduling to the load scheduling optimizability after the initial scheduling is taken as the optimizability ratio, which can be expressed as: .

[0052] The utilization rate change of each preset hardware is obtained based on the utilization rate before and after the initial scheduling. The process of obtaining the utilization rate change includes: for any hardware, subtracting the utilization rate before the initial scheduling from the utilization rate after the initial scheduling to obtain the increase in utilization, and dividing the increase in utilization by the utilization rate after the initial scheduling as the utilization rate change of that hardware. The utilization rate change can be expressed as: And calculate the average value of the utilization rate change, which can be expressed as: .

[0053] The utilization improvement is obtained based on the average of the optimizability ratio and the utilization rate change magnitude.

[0054] The utilization improvement can be expressed by the formula: ; Among them, the This indicates the optimizability of the load scheduling before the initial scheduling. Indicates the degree of optimizability of the load scheduling after the initial scheduling, subscript and This is only used to distinguish between before and after the initial scheduling. Indicates the first Within the first edge server The utilization rate of the first hardware unit after scheduling minus the first hardware unit The utilization rate of the k-th hardware within an edge server before scheduling, the Indicates the first Within the first edge server The utilization rate of each hardware component after scheduling, the This indicates the number of preset hardware components. This indicates that the request signal will be scheduled to the first... The improvement in the overall utilization of the cloud server after adding an edge server.

[0055] Figure 3 The flowchart below shows the process for obtaining load dispersion according to the third embodiment of the present invention. The process for obtaining load dispersion includes: S301. Calculate the absolute value of the difference between the utilization rate of the target hardware and the average utilization rate, wherein the target hardware is any of the preset hardware, and the average utilization rate is the average value of the utilization rates of the target hardware in the edge server.

[0056] When a single server processes multiple requests in parallel, parallel processing does not actually process multiple requests simultaneously. Therefore, parallel processing does not guarantee acceleration. Its efficiency will be much lower than the speed at which multiple servers process a single request. Therefore, judging the overall utilization of cloud servers solely based on the average of each edge server is too one-sided. Further evaluation is needed, taking into account the load balancing of each server.

[0057] A well-balanced server load should be close to the overall load average, and this absolute value can be expressed as: , wherein Indicates the first The utilization rate of target hardware in each edge server, the average utilization rate Indicates the average utilization of the target hardware, subscript It has no real meaning and is only used to describe the target hardware.

[0058] S302. Calculate the coefficient of variation of the utilization rate of the target hardware.

[0059] The coefficient of variation can be expressed as follows: , wherein The standard deviation of the utilization rate of the target hardware is represented by the This represents the average utilization rate of the target hardware.

[0060] S303. Calculate the load dispersion based on the ratio between the absolute value and the utilization rate of the corresponding target hardware, and the coefficient of variation.

[0061] The load dispersion can be expressed by the formula: ; Among them, the Indicates the number of edge servers, the This represents the load dispersion corresponding to the target hardware.

[0062] Since the target hardware is any of the preset hardware, the preset hardware has One, therefore There are also indivual.

[0063] The Indicates the first The absolute value of the difference between the utilization rate of the target hardware and the average utilization rate in each edge server is compared with the above. The ratio represents the first The difference between the utilization rate of the target hardware and the average utilization rate in each edge server is the proportion of its original utilization rate. The larger this value, the better the utilization rate of the target hardware. The higher the dispersion of an edge server relative to the whole, the better.

[0064] The The larger the value, the greater the load dispersion.

[0065] Figure 4 The flowchart illustrates the suitability acquisition process provided in the fourth embodiment of the present invention. The suitability acquisition process includes: S401. Calculate the ratio between the CPU utilization of the preferred server and the CPU utilization of the alternative server, and record it as the CPU utilization ratio.

[0066] The CPU utilization ratio can be expressed as: , wherein This indicates the CPU utilization of the preferred server. This indicates the CPU utilization rate of the alternative servers.

[0067] S402. Obtain the first distance between the preferred server and the user terminal and the second distance between the alternative server and the user terminal.

[0068] The first distance between the preferred server and the user terminal can be used as... The second distance between the alternative server and the user terminal can be represented by... To express.

[0069] S403. The suitability is obtained based on the CPU utilization ratio and the difference between the second distance and the first distance.

[0070] The suitability can be expressed by the formula: ; Among them, the This indicates the suitability of secondary scheduling to the b-th edge server, i.e., the alternative server. o is a preset parameter. In this embodiment of the invention, the value of the preset parameter can be 0.01. In other embodiments, the implementer can adjust the value according to the actual situation. The purpose is to avoid the situation where the denominator is 0, which would make the ratio meaningless.

[0071] Figure 5 This is a schematic diagram of a distributed cloud server load balancing optimization system based on edge computing provided in the fifth embodiment of the present invention. The system includes: The receiving module 501 is used to receive request signals sent by the user terminal; The initial scheduling module 502 is used to initially schedule the request signal to the preferred server and obtain the first utilization improvement of the overall cloud server after the initial scheduling. The preferred server is the edge server with the shortest distance to the user terminal. The secondary scheduling module 503 is used to reschedule the request signal to a candidate server and obtain the suitability of rescheduling to the candidate server and the second utilization improvement of the secondary scheduling on the overall cloud server. The candidate server is any edge server other than the preferred server. The final scheduling determination module 504 is used to obtain a scheduling match degree based on the suitability, the first utilization improvement degree and the second utilization improvement degree. If the scheduling match degree with the largest value is a non-negative number, then the candidate server corresponding to it is the final server; otherwise, the preferred server is the final server, and the final server serves the request signal.

[0072] The technical features and effects of the distributed cloud server load balancing optimization system based on edge computing proposed in this embodiment of the invention are the same as those of the method proposed in this embodiment of the invention, and will not be repeated here.

[0073] The present invention has the following beneficial effects: First, a request signal sent by the user terminal is received. This request signal forms the basis for subsequent analysis.

[0074] Secondly, the request signal is initially scheduled to the preferred server, and the initial utilization improvement of the overall cloud server after the initial scheduling is obtained. The preferred server is the edge server with the shortest distance to the user terminal. The higher the initial utilization improvement, the higher the overall utilization improvement of the cloud server after the initial scheduling.

[0075] Furthermore, the request signal is rescheduled to a candidate server, and the suitability of this rescheduling and the second utilization improvement of the secondary scheduling on the overall cloud server are obtained. The candidate server is any edge server other than the primary server. A higher second utilization improvement indicates a higher overall utilization improvement of the cloud server due to the secondary scheduling.

[0076] Finally, a scheduling match score is obtained based on the suitability, the first utilization improvement score, and the second utilization improvement score. If the scheduling match score with the highest value is non-negative, the corresponding candidate server is the final server; otherwise, the preferred server is the final server, and the final server serves the request signal. When the scheduling match score with the highest value is non-negative, it indicates that the secondary scheduling has a greater improvement in the overall utilization of the cloud server compared to the initial scheduling. The higher the non-negative value of the scheduling match score with the highest value, the better the effect of selecting the candidate server corresponding to the scheduling match score with the highest value to serve the request signal.

[0077] In summary, by using this invention, the matching degree between user requests and each edge server can be comprehensively considered, load scheduling can be carried out reasonably, and the overall resource utilization of cloud servers can be improved.

[0078] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0079] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A distributed cloud server load balancing optimization method based on edge computing, characterized in that, The method comprises: receiving a request signal sent by a user terminal; initially scheduling the request signal to a preferred server and obtaining a first utilization improvement degree of the overall cloud server after initial scheduling, the preferred server being an edge server with the shortest distance value from the user terminal; secondarily scheduling the request signal to an alternative server and obtaining a suitability degree of the alternative server after secondary scheduling and a second utilization improvement degree of the overall cloud server after secondary scheduling, the alternative server being any edge server other than the preferred server; obtaining a scheduling matching degree based on the suitability degree, the first utilization improvement degree and the second utilization improvement degree, and if the scheduling matching degree with the maximum value is a non-negative number, the corresponding alternative server is a final server, otherwise, the preferred server is the final server, and the final server serves the request signal.

2. The edge-computing-based distributed cloud server load balancing optimization method of claim 1, wherein, The process of obtaining the utilization improvement degree comprises: obtaining a load dispersion degree of each preset hardware in the edge server, and obtaining a load scheduling optimization degree of the overall cloud server based on the load dispersion degree; obtaining the load scheduling optimization degree before and after initial scheduling and the utilization rate of each preset hardware to calculate the utilization improvement degree.

3. The edge-computing-based distributed cloud server load balancing optimization method of claim 2, wherein, The process of obtaining the load dispersion degree comprises: calculating the absolute value of the difference between the utilization rate of a target hardware and the average utilization rate of the target hardware in the edge server, the target hardware being any of the preset hardware; calculating the coefficient of variation of the utilization rate of the target hardware; calculating the load dispersion degree based on the ratio between the absolute value and the utilization rate of the corresponding target hardware and the coefficient of variation.

4. The edge-computing-based distributed cloud server load balancing optimization method of claim 2, wherein, The process of obtaining the load scheduling optimization degree comprises: normalizing the average value of the load dispersion degrees of the preset hardware to obtain the load scheduling optimization degree.

5. The edge-computing-based distributed cloud server load balancing optimization method of claim 2, wherein, The process of obtaining the utilization improvement degree further comprises: taking the ratio of the load scheduling optimization degree before initial scheduling to the load scheduling optimization degree after initial scheduling as an optimization degree ratio; obtaining the utilization rate change range of each preset hardware according to the utilization rates of the preset hardware before and after initial scheduling, and calculating the average value of the utilization rate change range; obtaining the utilization improvement degree based on the optimization degree ratio and the average value of the utilization rate change range.

6. The edge-computing-based distributed cloud server load balancing optimization method of claim 5, wherein, The process of obtaining the utilization rate change range comprises: for any hardware, subtracting the utilization rate before initial scheduling from the utilization rate after initial scheduling to obtain a utilization rate increase, and dividing the utilization rate increase by the utilization rate after initial scheduling to obtain the utilization rate change range corresponding to the hardware.

7. The edge-computing-based distributed cloud server load balancing optimization method of claim 1, wherein, The process of obtaining the suitability degree comprises: calculating the ratio between the CPU utilization rate of the preferred server and the CPU utilization rate of the alternative server as a CPU utilization rate ratio; obtaining a first distance between the preferred server and the user terminal and a second distance between the alternative server and the user terminal; The fitness is obtained based on the CPU utilization ratio and a difference between the second distance and the first distance.

8. The edge-computing-based distributed cloud server load balancing optimization method of claim 1, wherein, The preset hardware comprises a CPU, a memory and an IO interface.

9. The edge-computing-based distributed cloud server load balancing optimization method of claim 1, wherein, The method further comprises: before receiving the request signal sent by the user terminal, initializing the cloud server configuration according to user demand, deploying the base station to form the wireless access network, connecting the user terminal through the communication network equipment, deploying the edge server at the network edge side to construct the edge computing system.

10. A distributed cloud server load balancing optimization system based on edge computing, characterized in that, The system comprises: a receiving module configured to receive a request signal sent by a user terminal; a first scheduling module configured to initially schedule the request signal to a preferred server and obtain a first utilization rate improvement degree of the overall cloud server after the initial scheduling, the preferred server being an edge server with the shortest distance value from the user terminal; a second scheduling module configured to secondarily schedule the request signal to a backup server and obtain a fitness of the second scheduling to the backup server and a second utilization rate improvement degree of the overall cloud server after the secondary scheduling, the backup server being any edge server other than the preferred server; a final scheduling determination module configured to obtain a scheduling matching degree based on the fitness, the first utilization rate improvement degree and the second utilization rate improvement degree, and if the scheduling matching degree of the maximum one is a non-negative number, the corresponding backup server is a final server, otherwise, the preferred server is the final server, and the final server serves the request signal.

Images