WO2018170732A1

WO2018170732A1 - Method and device for service deployment under edge cloud architecture

Info

Publication number: WO2018170732A1
Application number: PCT/CN2017/077510
Authority: WO
Inventors: 徐琳; 胡俊; 于正元
Original assignee: 华为技术有限公司
Priority date: 2017-03-21
Filing date: 2017-03-21
Publication date: 2018-09-27

Abstract

A method and a device for service deployment under edge cloud architecture, for improving the resource utilization rate of a system. In an embodiment of the present application, according to the number of available resource instances corresponding to a first target node and information concerning the access amount of a service by users in each area among N areas, the number of resource instances that should be allocated to the service by the first target node is determined. Therefore, the number of resource instances allocated to the service can be more reasonably determined on the basis of a supply and demand relationship, thereby efficiently improving the resource utilization rate. Further, information concerning the access amount of a service by users in each area among the N areas is predicted, and the total communication delays required for users in the N areas to access the service after the service has been deployed on the node can be determined according to the information concerning the amount of access. Therefore, the solution of choosing, according to M total communication delays, the first target node on which the service is deployed can further reduce the communication delay of a user of each area among the N areas accessing the service.

Description

Method and device for service deployment under edge cloud architecture

Technical field

The embodiments of the present invention relate to the field of communications technologies, and in particular, to a method and apparatus for service deployment in an edge cloud architecture.

Background technique

With the advent of the 5G era, mobile communication networks have increased new demands such as massive connectivity and ultra-low access latency. The traditional centralized data center architecture is facing challenges, and the interests of operators will inevitably be affected. In order to reduce the service response delay and meet the massive connection requirements, the industry has expanded on the basis of traditional cloud computing systems and proposed an edge cloud architecture.

In the micro cloud architecture, various services are “sinked” from the data center to the micro cloud nodes, thereby dispersing the access traffic of mobile terminal users, reducing network congestion and shortening the access delay. In the Cloud Radio Access Network (Cloud RAN) architecture, the edge cloud platform may include network element services that are sinked to the edge by the Evolved Packet Core (EPC) module. It includes services that are “upward” by the indoor baseband processing unit (BBU), such as Packet Data Convergence Protocol (PDCP), Radio Resource Control (RRC), etc., and even some Three-party application services. However, no matter which form of edge cloud architecture, resources such as computing, storage, and bandwidth are limited in a single node, and all services cannot be deployed at the same time. Therefore, how to reasonably arrange the placement of services on each node and the allocation of resource instances becomes a top priority.

Currently, the service deployment solution under the edge cloud architecture has not appeared in the industry. Existing mainstream cloud platform systems provide custom resource configurations. Specifically, the user requests the number of resource instances for the service before running the resource instance of the cloud service. In order to protect the quality of their services, users in this solution often apply for the number of reserved resource instances too much, resulting in waste of resources.

Summary of the invention

The embodiments of the present application provide a method and apparatus for service deployment in an edge cloud architecture, which are used to improve resource utilization of the system.

In a first aspect, the embodiment of the present application provides a service deployment method in an edge cloud architecture. When the K services to be deployed are deployed to the M nodes, the service is performed for each of the K services to be deployed. According to the historical log information, the information about the user's access to the service in each of the N regions is estimated; N is an integer greater than or equal to 1; for each of the M nodes, the service deployment is determined according to the access information. After the node, the total communication delay required by the users in the N areas to access the service obtains M total communication delays; according to the M total communication delays, the first target to be deployed by the service is determined from the M nodes. a node; determining, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, the number of resource instances that the first target node should allocate for the service; A resource instance corresponding to the number of resource instances on the first target node is allocated to the service, where K and M are integers greater than or equal to 1.

Optionally, for each of the M nodes, determining, according to the visitor information, a total communication delay required for a user of the N areas to access the service after the service is deployed on the node, and obtaining M total communication delays, including : Performing each of the M nodes as the current node, respectively: performing: determining, according to the access information and the communication performance indicator of the current node, the basic communication when the user of the N areas accesses the service after the service is deployed on the current node If it is determined that the node currently deployed by the service is different from the current node, it is determined that the service is migrated from the currently deployed node to the current node. The migration communication delay corresponding to the process; according to the basic communication delay and the migration communication delay, the total communication delay required for the users of the N areas to access the service is determined.

Specifically, the current cycle needs to deploy K services on M nodes, but there may be cases where K services have been deployed, this time being redeployed. At this time, if the node currently deployed by the service is different from the current node, once the service is determined to be deployed on the current node, all data corresponding to the service needs to be migrated to the current node, and the process of migrating the service must take time. Therefore, in the embodiment of the present application, the basic communication delay after the service migration and the migration communication delay corresponding to the process of migrating the service from the currently deployed node to the current node are considered, and the total communication can be obtained more accurately and completely. Delay. Optionally, the sum of the basic communication delay and the migration communication delay is taken as the total communication delay.

In the embodiment of the present application, according to the M total communication delays, the first target node where the service should be deployed is determined from the M nodes, and optionally, the minimum total communication may be determined from the M total communication delays. The delay determines the node corresponding to the smallest total communication delay as the first target node corresponding to the service.

Optionally, determining, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, the number of resource instances that the first target node should allocate for the service, including Determining the consumption of the resource instance corresponding to the service according to the performance indicator of the service and the information of the user's access to the service in each of the N areas; the number of available resource instances corresponding to the first target node, and the resource corresponding to the service Instance consumption, determining the number of tasks that the first target node can concurrently process the service; determining the number of tasks according to the number of tasks of the first target node concurrently processing the service, and the user access information of the service in each of the N areas The number of resource instances that a target node should allocate for the service.

The number of resource instances allocated by the first target node for the service is determined according to the number of available resource instances corresponding to the first target node and the user access information of the service in each of the N regions. Therefore, the number of resource instances allocated for the service can be determined more reasonably based on the supply and demand relationship, thereby effectively improving resource utilization.

Optionally, determining, according to the number of tasks of the first target node concurrently processing the service, and the user access information of the service in each of the N areas, determining the number of resource instances that the first target node should allocate for the service, including Determining the area of the N areas with the largest amount of access information of the service as the target area; determining the ideal access amount information of the user of the target area to the service when the service is deployed on the first target node; The traffic information of the service, the number of tasks of the first target node concurrently processing the service, and the ideal traffic information of the user of the target area to the service determine the number of resource instances that the first target node should allocate for the service.

In the embodiment of the present application, the number of available resource instances corresponding to the first target node may be determined, and the number of resource instances on each node of the M nodes is not infinite, and the corresponding total number of resource instances is used. The total number of resource instances minus the number of resource instances that have been occupied, that is, the number of available resource instances. It can be seen that the relationship between the user access traffic and the resource consumption is considered in the embodiment of the present application, so that the number of resource instances can be allocated to the service more reasonably.

Optionally, determining, if the service is deployed on the first target node, the ideal access information of the user of the target area to the service, including: determining, if the service is deployed on the first target node, according to the target user access in the target area The traffic access information of the service obtains the ideal visitor information of the user in the target area; wherein, for each user in the target area, if the distance between the location of the user and the location of the first target node is preset Within the distance range, the user is determined to be the target user.

It can be seen that, in the embodiment of the present application, when determining the number of resource instances that the first target node should allocate for the service, the ideal access information of the user in the target area is also considered, so that the consideration is more comprehensive, thereby further improving. Resource utilization.

Optionally, determining, according to the user's access information of the service in the target area, the number of tasks of the first target node concurrently processing the service, and the ideal access information of the user of the target area to the service, determining that the first target node should be allocated for the service The number of resource instances includes: determining the user's access information to the service in the target area, the number of tasks of the first target node concurrent processing service, and the minimum value of the user's ideal access information to the service in the target area as: The number of resource instances that a target node should allocate for the service.

Optionally, after the determined resource instance corresponding to the number of resource instances on the first target node is allocated to the service, the method further includes: if there is a constraint corresponding to the service, and the constraint includes: deploying the service to the M nodes A second target node, wherein the second target node is different from the first target node, and the service is deployed on the second target node.

Optionally, the constraint further includes: allocating a target number of resource instances to the service on the second target node; and deploying the service on the second target node, including: if the number of available resource instances on the second target node is less than the target quantity, Executing: deleting the deployed service on the second target node and minimizing the number of allocated resource instances from the second target node until the number of available resource instances on the second target node is greater than the target number; The service is deployed on the node and a target number of resource instances are assigned to the service.

Optionally, after the K services to be deployed are deployed to the M nodes, the method further includes: deploying K services to be deployed to the M nodes and then K services according to the number of resource instances allocated by the K services. The corresponding user access information is used to obtain the system resource utilization rate corresponding to the K services; if it is determined that the system resource utilization is less than the preset resource utilization threshold, the K services are redeployed.

In a second aspect, an embodiment of the present application provides a service deployment apparatus in an edge cloud architecture, where a service deployment apparatus in an edge cloud architecture includes a memory and a processor, where: the memory is used to store an instruction; and the processor is configured to perform storage according to the memory. The instructions, when the processor executes the instructions stored in the memory, the service deployment apparatus under the edge cloud architecture is configured to perform the method of any of the first aspect or the first aspect.

In a third aspect, the embodiment of the present application provides a service deployment device in an edge cloud architecture, which is used to implement any one of the foregoing first aspect or the first aspect, including a corresponding functional module, which is respectively used to implement the foregoing method. The steps in .

In a fourth aspect, an embodiment of the present application provides a computer readable storage medium, where the computer readable storage medium stores instructions that, when run on a computer, cause the computer to perform the first aspect or any possible implementation of the first aspect. The method in the way.

In a fifth aspect, an embodiment of the present application provides a computer program product comprising instructions, when executed on a computer, causing a computer to perform the method of the first aspect or any possible implementation of the first aspect.

In the embodiment of the present application, the number of resource instances allocated by the first target node for the service is determined according to the number of available resource instances corresponding to the first target node and the user access information of the service in each of the N regions. . Therefore, the number of resource instances allocated for the service can be determined more reasonably based on the supply and demand relationship, thereby effectively improving resource utilization.

Further, in the embodiment of the present application, the information about the access of the user to the service in each of the N areas is estimated, and then the information required by the user to access the service in the N areas after the service is deployed may be determined according to the access information. The total communication delay is such that the communication delay of the user accessing the user in each of the N areas can be further reduced by selecting the first target node deployed by the service according to the M total communication delays.

DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments will be briefly described below.

1 is a schematic diagram of a system of a micro cloud architecture according to an embodiment of the present application;

2 is a schematic flowchart of a service deployment method in an edge cloud architecture according to an embodiment of the present disclosure;

FIG. 3 is a schematic flowchart of a service access method according to an embodiment of the present application;

FIG. 4 is a schematic diagram of a micro cloud architecture according to an embodiment of the present application;

FIG. 5 is a schematic diagram of another micro cloud architecture provided by an embodiment of the present application;

FIG. 6 is a schematic diagram of another micro cloud architecture according to an embodiment of the present application;

FIG. 7 is a schematic flowchart of a service deployment method in an edge cloud architecture according to an embodiment of the present disclosure;

FIG. 8 is a schematic flowchart of a service deployment method in an edge cloud architecture according to an embodiment of the present disclosure.

detailed description

In order to make the objects, technical solutions and beneficial effects of the present application more clear, the present application will be further described in detail below with reference to the accompanying drawings and embodiments.

FIG. 1 is a schematic diagram showing a system of a micro cloud architecture to which the embodiment of the present application is applied. As shown in FIG. 1 , the micro cloud architecture includes a data center 101 and a plurality of autonomous regions, which are an autonomous region 102, an autonomous region 103, and an autonomous region. 104. Each autonomous region corresponds to at least one micro cloud node, for example, the autonomous region 102 includes a micro cloud node 105, a micro cloud node 106, and a micro cloud node 107; the autonomous region 103 includes a micro cloud node 108, a micro cloud node 109, and a micro cloud node 110; The micro cloud node 111, the micro cloud node 112, and the micro cloud node 113 are included.

In the micro cloud architecture, various services sink from the data center to the micro cloud node, and users in each area can access the services that sink to the micro cloud nodes, and the mobile terminal users are dispersed through the micro cloud architecture, and the number of users is reduced. Network congestion and reduced access latency. The coverage of the micro cloud node includes many users, and the user can be geographically divided into different areas. For example, a province can be divided into one area. The user shown in FIG. 1 includes users of two areas, namely the user of the area 121 and the user of the area 122. A user in one area can access services deployed on multiple nodes. At a physical distance, some users in area 121 may be closest to micro cloud node 109, and other users in the area may be closest to micro cloud node 110. Assuming that the service that the user needs to access is deployed on both the micro cloud node 109 and the micro cloud node 110, the user in the area 121 closest to the micro cloud node 109 can access the service deployed on the micro cloud node 109, and is closest to the micro cloud node 110. The user can access the services deployed on the micro cloud node 110. It can be seen that the user can select the service on the micro cloud node closer to himself, thereby shortening the delay of accessing the service.

Based on the system architecture shown in FIG. 1 , FIG. 2 exemplarily shows a schematic flowchart of a method for deploying services in an edge cloud architecture according to an embodiment of the present application. As shown in FIG. 2 , the method is applicable to K to be deployed. The application is deployed to the application scenarios of M nodes. The application scenario in this embodiment may be further divided into two types. The first type, the M nodes may refer to the M autonomous regions shown in FIG. 1, that is, the K services are deployed to the M autonomous regions, and the M is The number of available resource instances for each node in the node specifically refers to the number of available resource instances of the autonomous region indicated by the node. The second type of M nodes may refer to the M micro cloud nodes in FIG. 1 , and specifically refers to deploying K services to M micro cloud nodes, and the number of available resource instances of each node in the M nodes at this time. Specifically, it refers to the number of available resource instances on the micro cloud node indicated by the node. The solution provided by the embodiment of the present application can also be applied to application scenarios of other service deployments. In the embodiment of the present application, K is an integer greater than or equal to 1; and M is an integer greater than or equal to 1. The service in the embodiment of the present application may include a functional unit of the service refers to an application, such as a Packet Data Convergence Protocol (PDCP) and a Radio Resource Control (Radio Resource) in a Radio Access Network (RAN) application. Control, RRC) are all services; resource instances may include virtual machines that provide computing services.

The service deployment method of the edge cloud architecture provided by the embodiment of the present application may be performed by the service deployment apparatus. Optionally, the service deployment apparatus may periodically deploy K services. An alternative implementation is that K services can be deployed to M nodes periodically. Or, each time K services are deployed, the nodes to be deployed may be re-determined, that is, each time K services are deployed, the nodes to be deployed corresponding to the K services may be different, partially different, or the same.

Another optional implementation is that after the K services to be deployed are deployed to the M nodes, the method further includes: deploying the K services to be deployed to the M services according to the number of resource instances allocated by the K services. If the system resource utilization is less than the preset resource utilization threshold, the K services are redeployed. Optionally, each time the K services are deployed, the nodes to be deployed corresponding to the K services may be different, partially different, or the same. Specifically, according to the user access amount information corresponding to the K services after the K services to be deployed are deployed to the M nodes, the actual resource consumption corresponding to the K services may be determined, and then the resource instances allocated according to the K services are determined. The amount and actual resource consumption determine the system resource utilization. The resource utilization threshold can be set according to the specific scenario and is an empirical value. In this embodiment, the solution can flexibly adjust the deployment of the service multiple times, thereby improving resource utilization of the deployed service.

Optionally, the method for determining the actual resource consumption corresponding to the K services according to the user access amount information corresponding to the K services after the K services to be deployed are deployed to the M nodes may be implemented in multiple manners, for example, a regression method may be utilized. , predict resource consumption by user traffic. An optional implementation manner is provided in the embodiment of the present application for predicting actual resource consumption corresponding to future K services. Optionally, the actual resource consumption corresponding to the K services is equal to a0, the product obtained by multiplying the user access information corresponding to the K services, and the product obtained by multiplying the service level agreement (SLA) constraint by a2. The sum of the three. Alternatively, a0, a1, and a2 are coefficients that can be found by linear fitting.

Optionally, the system resource utilization is determined according to the number of resource instances allocated by the K services and the actual resource consumption. The embodiment of the present application provides an optional implementation manner for determining system resource utilization. Rate, specifically, the system resource utilization may include the resource utilization of the CPU, and the system resource utilization is equal to the total amount of used CPU resources on all virtual machines on the M micro cloud nodes and all virtual machine ratings on the M micro cloud nodes. The product of the total amount of CPU resources, and the total amount of used CPU resources on all virtual machines on the M micro-cloud nodes can be determined by the actual resource consumption corresponding to the predicted future K services. The resource utilization calculations of other dimensions are similar. No longer detailed.

The service deployment method in the embodiment of the present application may be performed by the service deployment apparatus. Specifically, the method includes: performing, for each service of the K services to be deployed, the following:

Step 201: The service deployment device estimates the access information of the user in each of the N areas according to the historical log information; N is an integer greater than or equal to 1;

Step 202: The service deployment device determines, for each of the M nodes, the total communication delay required for the users in the N areas to access the service after the service is deployed on the node, and obtains M total communications. Delay

Step 203: The service deployment device determines, according to the M total communication delays, the first target node that the service should be deployed from the M nodes.

Step 204: The service deployment device determines, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, the resource instance that the first target node should allocate for the service. quantity;

Step 205: The service deployment device divides the determined resource instance corresponding to the number of resource instances on the first target node. Provision this service.

In the foregoing step 201, the historical access record is analyzed according to various analysis tools, so that the user's access information of the service to each service in each of the N regions can be obtained, and further, the historical tool-based access amount can be obtained through the analysis tool. The information estimates the amount of access to the service by users in each of the N regions in the future, such as analysis by a time series analysis tool. The traffic information of the user of each of the N regions may include an estimated traffic distribution of the future users, specifically, the amount of access to the service in each of the N regions in the future, or Say the amount of access to the service by users in the area at a certain moment in the future. The traffic information may also include information such as future trends of the traffic, access time, and the like.

In the foregoing step 202, optionally, the communication model is also obtained in the embodiment of the present application, and the communication model may include a bandwidth description, a communication delay, and the like, and a communication performance indicator between each node of the M nodes, according to the communication performance. The metric determines the basic communication delay for the user to access the service. The node specification parameters of each node of the M nodes are also obtained, and the specification parameters of the node may include information such as a central processing unit (CPU), a memory, a hard disk, a number of network cards, and a network card bandwidth of the node. Based on the above information, the migration communication delay corresponding to the process of migrating the service from the currently deployed node to the current node can be determined.

Optionally, for each of the M nodes, the total communication delay required for the users in the N areas to access the service after the service is deployed on the node is determined according to the access information, and the total communication delay is obtained. The method includes: using each of the M nodes as the current node, performing: determining, according to the access information and the communication performance indicator of the current node, that the user of the N areas accesses the service after the service is deployed in the current node The basic communication delay; if it is determined that the node currently deployed by the service is different from the current node, the migration communication delay corresponding to the process of migrating the service from the currently deployed node to the current node is determined; according to the basic communication Delay and migration communication delays determine the total communication delay required for users in N areas to access the service.

In the embodiment of the present application, the basic communication delay and the migration communication delay caused by the service migration are determined by using the user access amount information, and then the service station is determined according to the basic communication delay and the migration communication delay caused by the service migration. The nodes that are deployed so that the total communication latency after the service is deployed is smaller.

In the embodiment of the present application, if it is determined that the node currently deployed by the service is the same as the current node, or the service has not been deployed, determining the total communication required for the users of the N areas to access the service according to the basic communication delay Delay. That is to say, when the node currently deployed by the service is the same as the current node, or the service has not been deployed, the deployment of the service can be performed only considering the basic communication delay.

Optionally, in the embodiment of the present application, the service may be deployed on the node in multiple representation manners. For example, for each node on the M nodes, the service deployment may be represented by 1 on the node, and the service is not deployed in the node. 0 table available on the node Show.

The solution provided in the foregoing step 204 determines, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, the resource that the first target node should allocate for the service. The number of instances. Therefore, the number of resource instances allocated for the service can be determined more reasonably based on the supply and demand relationship, thereby effectively improving resource utilization.

Optionally, determining the number of resource instances that the first target node should allocate for the service according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions. The method includes: determining, according to the performance indicator of the service, and the user access information of the service in each of the N areas, the consumption of the resource instance corresponding to the service; according to the number of available resource instances corresponding to the first target node, And the resource instance consumption corresponding to the service, determining the number of tasks that the first target node can concurrently process the service; the number of tasks that process the service concurrently according to the first target node, and the users of each of the N regions The traffic information of the service determines the number of resource instances that the first target node should allocate for the service.

Optionally, in the embodiment of the present application, according to the number of available resource instances corresponding to the first target node and the consumption of the resource instance corresponding to the service, determining that the number of tasks that the first target node can concurrently process the service has multiple implementation manners The present application provides an alternative implementation manner. Assume that the amount of idle resources on a node is R, and the user model can be processed by the performance model function to obtain the number of user accesses (ie, the number of tasks) that the node can currently process. Since the resources are multi-dimensional, each dimension repeats the above operations, and finally obtains an estimated number of tasks that can be processed. Alternatively, the smallest one can be selected as the number of tasks that the node can currently process concurrently.

Optionally, the resource instance consumption can be determined by using a certain analysis tool. For example, the test data of the service submitted by the user and the log information of the service during the running process can be analyzed by using the mining tool and the regression analysis tool. Get the service performance model. The service performance model is specifically a mapping between the service performance indicator and the traffic information and the resource instance consumption. For example, the resource instance consumption can be obtained according to the service performance model, the service performance indicator and the traffic information. The performance indicators of the service may include delay information, packet loss rate, and the like of the service.

Further, optionally, according to the number of tasks of the first target node concurrently processing the service, and the information about the access amount of the service of the user of each of the N areas, the resource that the first target node should allocate for the service is determined. The number of the instances includes: determining an area in the N areas that has the largest amount of access information for the service as the target area; determining an ideal amount of access to the service by the user of the target area if the service is deployed on the first target node Information; determining, according to the user's access information of the service in the target area, the number of tasks that the first target node concurrently processes the service, and the ideal visitor information of the user of the target area to the service, determining that the first target node should be the The number of resource instances allocated by the service.

As an example, assume that the resources on each node of the M nodes are infinite, and K services are deployed on each node, and the number of resource instances allocated by each service is unlimited. At the time, for each of the K services, when the user of each of the N areas accesses the service, for each user of each of the N areas, the user accesses the node closest to himself. Service. The preset distance range can be set smaller. At this time, it is determined that the service is deployed on the first target node, and the user who is within the preset distance from the first target node accesses the first target node. The deployed service, the user whose distance from the first target node exceeds the preset distance does not access the service deployed on the first target node.

It can be seen that, in the embodiment of the present application, when determining the number of resource instances that the first target node should allocate for the service, the ideal access information of the user in the target area is also considered, so that the consideration is more comprehensive, thereby further improving resources. Utilization rate.

Optionally, the first target node is determined according to the information about the access of the service by the user of the target area, the number of tasks of the first target node concurrently processing the service, and the ideal visitor information of the user of the target area to the service. The number of resource instances allocated for the service includes: information on the amount of access of the user in the target area to the service, the number of tasks in which the first target node concurrently processes the service, and the ideal access amount information of the user in the target area to the service The minimum value is determined as the number of resource instances that the first target node should allocate for the service.

After the K services are deployed in the M nodes in the embodiment of the present application, some constraints may be considered, and then the deployment schemes of the K services may be fine-tuned according to the constraint conditions. Optionally, after the determined resource instance corresponding to the number of the resource instances on the first target node is allocated to the service, the method further includes: if the constraint corresponding to the service exists, and the constraint includes: deploying the service to the M A second target node in the node, wherein the second target node is different from the first target node, and the service is deployed on the second target node.

Optionally, the constraint further includes: allocating a target number of resource instances to the service on the second target node; and deploying the service on the second target node, including: if the number of available resource instances on the second target node is smaller than the target The quantity is executed by: deleting the service that has been deployed on the second target node and minimizing the number of allocated resource instances from the second target node until the number of available resource instances on the second target node is greater than the target quantity; The service is deployed on the second target node and a target number of resource instances are assigned to the service.

For example, service a is assigned to node 1, the constraint is to deploy service a to node 2, and service a is assigned 100 resource instances, and current node 2 deploys three services, namely service b and service. c and service d, service b allocates 120 resource instances, service c allocates 90 resource instances, service d allocates 60 resource instances, and node 2 has 20 remaining resource instances. In the embodiment of the present application, the service d and the service c on the node 2 are deleted, then the service a is deployed on the node 2, and 100 resource instances are allocated for the service a. Optionally, after the service d and the service c are deleted, the service d and the service c can be reported to the data center, and arranged by the data center, for example, the next deployment and redeployment service d and service c, or the service d and the service cDeploy to a few nodes and so on.

FIG. 3 is a schematic flowchart diagram of a service access method provided by an embodiment of the present application. As shown in FIG. 3, the method includes:

In step 301, the base station forwards the user's service access request (Request for transmitting) to the scheduling center (Scheduler) in the autonomous area head node, and then performs step 302.

Step 302: The scheduler queries the search resource deploy file, and then performs step 303 or step 305 according to the situation.

Step 303: If the scheduling center queries the service resource list to the micro-cloud node (Cloudlet) in which the service is deployed, the service access request of the user is forwarded to the corresponding micro-cloud node, and then step 304 is performed.

Step 304: After processing the user access request, the micro cloud node returns a result service response (Response for services) to the user, and ends.

Step 305: If the scheduling center queries from the service resource list that there is no micro cloud node in the autonomous region where the service is deployed, the dispatch center sends the cross-autonomous region to the Autonomous Region Resource Manager (AZ RM). Request for services across AZ, and then step 306 is performed.

Step 306: The micro cloud autonomous region resource manager sends a cross-autonomous service access request to the search engine (SearchEngine) of the data center, and then step 307 is performed.

Step 307: The search engine searches for an autonomous area (Search available AZ) in which the service is deployed, selects an optimal autonomous area, and sends a request for available cloudlet to the resource manager of the optimal autonomous area. Go to step 308.

Step 308, the optimal autonomous region resource manager searches for the optimal micro cloud node (Search available Cloudlet) deployed with the service, and returns the information of the optimal micro cloud node to the search engine, and then performs step 309.

Step 309: The search engine returns information of the optimal micro cloud node to the dispatch center that originally issued the service access request by using the micro cloud autonomous region resource manager, and then step 310 is performed.

Step 310: The scheduling center sends a user service access request to the optimal micro cloud node, and then step 311 is performed.

Step 311: The optimal micro cloud node returns the result service access response to the user after processing the user access request.

Based on the above, the following examples are also provided in the embodiments of the present application for introducing the present application.

Example one

FIG. 4 is a schematic diagram showing a micro cloud architecture provided by an embodiment of the present application. As shown in FIG. 4, the data center and the micro cloud node are included. Under the micro cloud architecture, the service deployment device can be integrated in the autonomous region resource manager of the data center and the autonomous region head node.

At the time of initial deployment:

First, after the user submits the service to be deployed to the deployment center service 401 (Deploy Center) of the data center 400, the deployment center service 401 forwards the service to be deployed to the first resource management service 402 of the data center 400 (Resource Mgr Service). ).

Second, the load analysis module 403 in the first resource management service 402 of the data center 400 models the performance of the service. In addition, the first resource management service 402 will also query the resource usage of each autonomous region, execute a service deployment algorithm, obtain a service deployment and load distribution scheme, and notify the deployment center service 401 of the result.

Third, the deployment center service 401 sends the service to the deployment agent service 407 of the corresponding micro cloud node for further deployment according to the deployment scenario.

Fourth, after receiving the service deployment request, the second resource management service 509 of the autonomous region head node in the micro cloud node 410 first queries the resource status of each node in the autonomous region through the resource proxy service 408, and executes the service deployment algorithm to obtain the service. Deploy and load distribution scenarios and inform the deployment center service 401 of the results.

When data center and autonomous service deployments are dynamically adjusted:

First, the first resource management service 402 processes the data of the monitoring service 405 of the analysis monitoring system, and when the system resource utilization is found to be below a certain threshold, initiates a service deployment adjustment request. Alternatively, the first resource management service 401 performs service deployment adjustment according to a clock set in advance.

Second, the first resource management service 402 can obtain a user access request record from the log service 406, and predict the distribution of user access requests in the future period through the analysis module.

Third, the first resource management service 402 derives service deployment and load through the deployment decision module 404 of the service deployment. The plan is assigned and the results are communicated to the first resource management service 402 of the data center.

In the above example, the load analysis module 403 can be respectively configured on the first resource management service 402 of the data center 400 and the second resource management service 409 of the micro cloud node 410, and the deployment decision module 404 can be respectively configured in the data center 400. A resource management service 402 and a second resource management service 409 of the micro cloud node 410. The service deployment device in the above may include the load analysis module 403 and the deployment decision module 404 in this example.

Example two

FIG. 5 is a schematic diagram showing a micro cloud architecture provided by an embodiment of the present application. As shown in FIG. 5, the data center and the micro cloud node are included. Under the micro cloud architecture, the service deployment appliance can be integrated into the Deploy Center Service in the data center.

At the time of initial deployment:

First, after the user submits the service to be deployed to the deployment center service 502 of the data center 500, the deployment center service 502 applies to the first resource management service 501 for the resource running service test case, and then performs performance analysis on the service through the load analysis module 503. mold.

Second, the deployment center service 502 obtains the micro-cloud nodes through the first resource management service 501 of the data center 500, the second resource management service 509 of the autonomous region head node in the micro-cloud node 510, and the resource proxy service 508 of the micro-cloud node. Resource status.

Third, the deployment center service 502 executes the deployment decision module 504 according to the predicted load distribution and the service performance model and the node resource status obtained in the first two steps, and obtains a solution for service deployment and load distribution.

Fourth, the deployment center service 502 deploys the service deployment plan to the deployment agent service 507 of each node for deployment.

The data center and the autonomous region service deployment are dynamically adjusted, similar to the first example, and will not be described here. For example, the first resource management service 501 can process the data of the monitoring service 405 of the analysis monitoring system, and initiate a service deployment adjustment request when the system resource utilization is found to be below a certain threshold. Alternatively, the first resource management service 501 performs service deployment adjustment according to a clock set in advance. The first resource management service 501 can obtain a user access request record from the log service 406, and the analysis module predicts a user access request distribution for a future period of time.

Example three

FIG. 6 is a schematic diagram showing a structure of a Cloud Radio Access Network (CloudRAN) provided by an embodiment of the present application. As shown in FIG. 6, the data center and the micro cloud node are included. Under the Cloud Radio Access Network (CloudRAN) architecture, the service deployment device is in the centralized data center platform layer management node.

At the time of initial deployment:

First, the user submits a service deployment request to the first deployment management and software management service 601. The first deployment management and software management service 601 applies to the resource management service 602 for the resource running service test case, and performs performance analysis on the service through the load analysis module 613. mold.

Second, the load analysis module 613 acquires the resource status of each micro cloud node through the resource management service 602. Combined with the predicted load distribution and service performance model, the deployment decision module 614 is executed to obtain a solution for service deployment and load distribution.

Third, the first deployment management and software management service 601 sends the service deployment plan to the second deployment management and software management service 603 of each of the aggregation nodes 604 for deployment.

Two algorithms can be provided in the embodiment of the present application. One algorithm can be called one-shot esds (enhanced service deployment solution), and the other can be called dynamic esds. One-shot esds can only consider the basic communication cost. The dynamic esds considers the cost of data transmission delay caused by basic communication delay and application service migration. From the perspective of delay, since the one-shot esds and dynamic esds algorithms consider the performance model of the services in the micro cloud nodes, the applications can be deployed as close as possible to the micro cloud nodes close to the users. Further, optionally, dynamic esds defaults to all nodes with unlimited resources, and only uses transmission delay as an optimization target, but in practice, due to limited node resources, task queuing is introduced, and additional waiting delay is introduced. The one-shot esds takes into account the resource limitations of the nodes, performs reasonable optimization, reduces the waiting delay, and makes the overall delay low. From the perspective of resource utilization, one-shot esds and dynamic esds flexibly change according to load changes, improving resource utilization.

The method provided by the embodiment of the present application is not limited to the service type and the optimization indicator, and has certain versatility and scalability. Specifically, for different services or new resource types that need to be considered, we can add appropriate analysis tools and algorithms in the analysis module to reasonably model the service and obtain an effective performance model. In addition, when there are new optimization needs, such as reducing the packet loss rate, we can achieve the new optimization goal by modifying the description of the corresponding utility benefit U.

FIG. 7 is a schematic structural diagram of a service deployment apparatus in an edge cloud architecture provided by an embodiment of the present application.

Based on the same concept, the embodiment of the present application provides a service deployment apparatus in an edge cloud architecture, which is used to execute the foregoing method flow. As shown in FIG. 7, the service deployment device 700 includes a load analysis module 701 and a deployment decision module 702. The load analysis module 701 can be the load analysis module 403 in FIG. 4, the load analysis module 503 in FIG. 5, or the above figure. In the load analysis module 613, the deployment decision module 702 can be the deployment decision module 404 in FIG. 4, the deployment decision module 504 in FIG. 5, or the deployment decision module 614 in FIG. 6 above.

The service deployment device 700 in the edge cloud architecture is applicable to an application scenario in which K services to be deployed are deployed to M nodes; wherein K and M are integers greater than or equal to 1.

The load analysis module 701 is configured to estimate each of the N regions according to the historical log information for each of the K services to be deployed when the K services to be deployed are deployed to the M nodes. User access information to the service; N is an integer greater than or equal to 1; for each of the M nodes, the total communication required for the user to access the service in the N areas after the service is deployed is determined according to the access information Delay, get M total communication delays.

The deployment decision module 702 is configured to determine, according to the M total communication delays, the first target node that the service should be deployed from the M nodes; the number of available resource instances corresponding to the first target node, and each of the N regions The user's access to the service information of the area determines the number of resource instances that the first target node should allocate for the service; and allocates the resource instance corresponding to the determined number of resource instances on the first target node to the service.

Optionally, the load analysis module 701 determines, for each of the M nodes, the total communication delay required for the users in the N areas to access the service after the service is deployed on the node, and obtains M totals. The communication time delay is specifically used to: each node of the M nodes is respectively used as the current node, and is executed according to the access amount information and the current section. The communication performance indicator of the point determines the basic communication delay of the user accessing the service in the N areas after the service is deployed in the current node; if it is determined that the node currently deployed by the service is different from the current node, it is determined that the service is from the current The migration communication delay corresponding to the process in which the deployed node migrates to the current node; according to the basic communication delay and the migration communication delay, the total communication delay required for the users of the N areas to access the service is determined.

Optionally, the deployment decision module 702 determines, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, the resource that the first target node should allocate for the service. The number of the instances is specifically used to: determine the consumption of the resource instance corresponding to the service according to the performance indicator of the service and the access information of the user in each of the N areas; and the available resources corresponding to the first target node The number of instances, and the resource instance consumption corresponding to the service, determine the number of tasks that the first target node can concurrently process the service; the number of tasks that concurrently process the service according to the first target node, and the user-to-service for each of the N regions The traffic information determines the number of resource instances that the first target node should allocate for the service.

Optionally, the deployment decision module 702 determines, according to the number of tasks of the first target node concurrently processing the service, and the user access information of the service in each of the N areas, the resource that the first target node should allocate for the service. The number of the instances is specifically used to: determine the area of the N areas with the largest amount of access information of the service as the target area; and determine the ideal amount of the user's access to the service when the service is deployed on the first target node. Information; determining, according to the user's access information of the service in the target area, the number of tasks of the first target node concurrently processing the service, and the ideal access information of the user of the target area to the service, determining the resource that the first target node should allocate for the service The number of instances.

Optionally, the deployment decision module 702 is specifically configured to: determine, if the service is deployed on the first target node, when the service is deployed on the first target node, when the service is deployed on the first target node, Obtaining ideal visitor information of the user in the target area according to the visitor information of the target user accessing the service in the target area; wherein, for each user in the target area, if the location of the user is located with the first target node If the distance between the locations is within a preset distance, it is determined that the user is the target user.

Optionally, the deployment decision module 702 determines the first information according to the user's access to the service information of the target area, the number of tasks of the first target node concurrently processing the service, and the ideal access information of the user of the target area to the service. When the target node should be the number of resource instances allocated to the service, it is specifically used to: the user's access information to the service in the target area, the number of tasks of the first target node concurrently processing the service, and the ideal access amount of the user in the target area to the service. The minimum value in the information is determined as the number of resource instances that the first target node should allocate for the service.

Optionally, after the deployment decision module 702 allocates the determined resource instance corresponding to the number of resource instances on the first target node to the service, the deployment decision module 702 is further configured to: if there is a constraint corresponding to the service, and the constraint includes: deploying the service to A second target node among the M nodes, wherein the second target node is different from the first target node, and the service is deployed on the second target node.

Optionally, the constraint further includes: allocating a target number of resource instances to the service on the second target node; and when the deployment decision module 702 deploys the service on the second target node, specifically: if the second target node is available If the number of resource instances is less than the target number, execute: deleting the deployed service on the second target node and minimizing the number of allocated resource instances from the second target node until the number of available resource instances on the second target node is greater than the target Up to the number; deploy the service on the second target node and assign a target number of resource instances to the service.

Optionally, the deployment decision module 702 deploys the K services to be deployed to the M nodes, and is further configured to: deploy the K services to be deployed to the M services according to the number of resource instances allocated by the K services. If the system resource utilization is less than the preset resource utilization threshold, the K services are redeployed.

It should be understood that the division of each unit above is only a division of a logical function, and the actual implementation may be integrated into one physical entity in whole or in part, or may be physically separated. In the embodiment of the present application, the load analysis module 701 and the deployment decision module 702 may be implemented by a processor. As shown in FIG. 8, the service deployment device 800 can include a processor 801 and a memory 803. The memory 803 can be used to store the code when the processor 801 executes the solution, and the code can be a program/code pre-installed by the service deployment device 800 at the factory.

FIG. 8 is a schematic structural diagram of a service deployment apparatus in an edge cloud architecture provided by an embodiment of the present application.

Based on the same concept, the embodiment of the present application provides a service deployment apparatus in an edge cloud architecture, which is used to execute the foregoing method flow. As shown in FIG. 8, the service deployment apparatus 800 includes a processor 801, a memory 803, and a communication interface 803; wherein the processor 801, the memory 803, and the communication interface 804 are connected to each other through a bus 802.

The bus 802 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus or the like. The bus can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is shown in Figure 8, but it does not mean that there is only one bus or one type of bus.

The memory 803 may include a volatile memory such as a random-access memory (RAM); the memory may also include a non-volatile memory such as a flash memory. A hard disk drive (HDD) or a solid-state drive (SSD); the memory 803 may also include a combination of the above types of memories.

The communication interface 804 can be a wired communication access port, a wireless communication interface, or a combination thereof, wherein the wired communication interface can be, for example, an Ethernet interface. The Ethernet interface can be an optical interface, an electrical interface, or a combination thereof. The wireless communication interface can be a WLAN interface.

The processor 801 can be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP. The processor 801 may further include a hardware chip. The above hardware chip may be an application-specific integrated circuit (ASIC), programmable logic Programmable logic device (PLD) or a combination thereof. The PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a general array logic (GAL), or any combination thereof.

The service deployment device 800 is applicable to an application scenario in which K services to be deployed are deployed to M nodes. Wherein K and M are integers greater than or equal to 1.

Optionally, the memory 803 can also be used to store computer program program instructions, and the processor 801 calls the computer program program instructions stored in the memory 803 to perform one or more steps in the embodiment shown in FIG. 2 and FIG. Or an optional embodiment thereof, such that the service deployment device 800 implements the functions of the service deployment device in the above method.

The processor 801 is configured to: according to the computer program instructions for executing the memory storage, when the processor 801 executes the computer program instructions stored in the memory, the service deployment apparatus is configured to: perform, for each service of the K services to be deployed, according to history The log information is used to estimate the access information of the user to the service in each of the N areas; N is an integer greater than or equal to 1; for each of the M nodes, the service is deployed on the node according to the access information. After that, the total communication delay required by the users in the N areas to access the service obtains M total communication delays; and based on the M total communication delays, the first target nodes to which the service should be deployed are determined from the M nodes; Determining, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, the number of resource instances that the first target node should allocate for the service; A resource instance corresponding to the number of resource instances on the target node is assigned to the service.

Optionally, the processor 801 determines, for each of the M nodes, the total communication delay required for the users in the N areas to access the service after the service is deployed on the node, according to the access information, and obtains M total communications. Time delay, specifically for: using each of the M nodes as the current node, performing: determining, according to the traffic information and the communication performance indicator of the current node, that the N service is deployed after the current node is deployed The basic communication delay of the user accessing the service; if it is determined that the node currently deployed by the service is different from the current node, the migration communication delay corresponding to the process of migrating the service from the currently deployed node to the current node is determined; The basic communication delay and the migration communication delay determine the total communication delay required for users in N areas to access the service.

In the embodiment of the present application, according to the M total communication delays, the first node that the service should be deployed is determined from the M nodes. The target node may optionally determine a minimum total communication delay from the M total communication delays, and determine the node corresponding to the smallest total communication delay as the first target node corresponding to the service.

Optionally, the processor 801 determines, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, the resource instance that the first target node should allocate for the service. The quantity is specifically used to: determine the consumption of the resource instance corresponding to the service according to the performance indicator of the service and the information of the user's access to the service in each of the N areas; and the available resource instance corresponding to the first target node The quantity, and the resource instance consumption corresponding to the service, determine the number of tasks that the first target node can concurrently process the service; the number of tasks that concurrently process the service according to the first target node, and the user-to-service of each of the N areas The traffic information determines the number of resource instances that the first target node should allocate for the service.

Optionally, the processor 801 determines, according to the number of tasks of the first target node concurrently processing the service, and the user access information of the service in each of the N areas, the resource instance that the first target node should allocate for the service. The quantity is specifically used to: determine the area of the N areas with the largest amount of access information of the service as the target area; and determine the ideal visitor information of the user of the target area to the service when the service is deployed on the first target node. Determining a resource instance that the first target node should allocate for the service according to the user's access information to the service in the target area, the number of tasks of the first target node concurrently processing the service, and the ideal access information of the user in the target area to the service quantity.

Optionally, the determining, by the processor 801, when the service is deployed on the first target node, when the user of the target area has the ideal access information of the service, the method is specifically configured to: determine, if the service is deployed on the first target node, according to The target user accesses the service access information in the target area to obtain the ideal visitor information of the user in the target area; wherein, for each user in the target area, if the user is located and the location of the first target node The distance between the preset distances determines that the user is the target user.

Optionally, the processor 801 determines the first target node according to the user's access information to the service according to the target area, the number of tasks of the first target node concurrently processing the service, and the ideal access information of the user of the target area to the service. When the number of resource instances to be allocated for the service is specifically used, the user's access information to the service in the target area, the number of tasks of the first target node concurrently processing the service, and the ideal access information of the user in the target area to the service are The minimum value is determined as the number of resource instances that the first target node should allocate for the service.

Optionally, after the processor 801 allocates the determined resource instance corresponding to the number of the resource instances on the first target node to the service, the processor 801 is further configured to: if there is a constraint corresponding to the service, and the constraint includes: deploying the service to the M A second target node among the nodes, wherein the second target node is different from the first target node, and the service is deployed on the second target node.

Optionally, the constraint further includes: allocating a target number of resource instances to the service on the second target node; and when the processor 801 deploys the service on the second target node, specifically: if the available resources on the second target node If the number of instances is smaller than the target quantity, execute: deleting the deployed service on the second target node and minimizing the number of allocated resource instances from the second target node until the number of available resource instances on the second target node is greater than the target number So far; deploy the service on the second target node and assign a target number of resource instances to the service.

Optionally, after the processor 801 deploys the K services to be deployed to the M nodes, the processor 801 is further configured to: deploy the K services to be deployed to the M nodes according to the number of resource instances allocated by the K services. Then, the user access amount information corresponding to the K services is used to obtain the system resource utilization rate corresponding to the K services; if it is determined that the system resource utilization is less than the preset resource utilization threshold, the K services are redeployed.

Those skilled in the art will appreciate that embodiments of the present application can be provided as a method, system, or computer program product. Therefore, the embodiments of the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Moreover, embodiments of the present application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) including computer usable program code.

Embodiments of the present application are described with reference to flowchart illustrations and/or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present application. It will be understood that each flow and/or block of the flowchart illustrations and/or FIG. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing device to produce a machine for the execution of instructions for execution by a processor of a computer or other programmable data processing device. Means for implementing the functions specified in one or more of the flow or in a block or blocks of the flow chart.

The computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device. The apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.

These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device. The instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.

It is apparent that those skilled in the art can make various modifications and variations to the embodiments of the present application without departing from the spirit and scope of the application. Thus, it is intended that the present invention cover the modifications and variations of the embodiments of the present invention.

The instructions executed above provide steps for implementing the functions specified in one or more blocks of the flowchart or in a block or blocks of the flowchart.

Claims

A method for service deployment in an edge cloud architecture is characterized in that, when deploying K services to be deployed to M nodes, for each of the K services to be deployed, perform:

Obtaining, according to the historical log information, information about the amount of access of the user in each of the N areas to the service; the N is an integer greater than or equal to 1;

Determining, for each of the M nodes, a total communication delay required for a user of the N areas to access the service after the service is deployed in the node according to the access information, M total communication delays;

Determining, from the M nodes, a first target node that the service should be deployed according to the M total communication delays;

Determining, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, that the first target node should be allocated for the service The number of resource instances;

Allocating the determined resource instance corresponding to the quantity of the resource instances on the first target node to the service;

The K and M are integers greater than or equal to 1.
The method according to claim 1, wherein, for each of the M nodes, determining, according to the access amount information, that the service is deployed in the node, the users of the N areas The total communication delay required to access the service, resulting in M total communication delays, including:

Performing each of the M nodes as the current node, respectively, performs:

Determining, according to the access information and the communication performance indicator of the current node, a basic communication delay of the user of the N areas accessing the service after the service is deployed in the current node;

If it is determined that the node currently deployed by the service is different from the current node, determining a migration communication delay corresponding to a process of migrating the service from the currently deployed node to the current node;

And determining, according to the basic communication delay and the migration communication delay, a total communication delay required for a user of the N areas to access the service.
The method according to claim 1 or 2, wherein, according to the number of available resource instances corresponding to the first target node, and the information on the amount of access of the user of each of the N regions to the service, Determining the number of resource instances that the first target node should allocate for the service, including:

Determining, according to the performance indicator of the service, the amount of access to the service by the user of each of the N areas, the resource instance consumption corresponding to the service;

Determining, according to the number of available resource instances corresponding to the first target node, and the resource instance consumption corresponding to the service, the number of tasks that the first target node can concurrently process the service;

Determining that the first target node should be the service according to the number of tasks of the first target node concurrently processing the service, and the user access information of the service of each of the N areas The number of resource instances assigned.
The method according to claim 3, wherein the number of tasks for processing the service concurrently according to the first target node, and the amount of access information of the user of each of the N regions to the service, Determining the number of resource instances that the first target node should allocate for the service, including:

Determining, as the target area, an area in the N areas that has the largest amount of access information of the service;

Determining an ideal amount of visitor information of the user of the target area to the service if the service is deployed on the first target node;

Determining, according to the information about the access of the service by the user of the target area, the number of tasks of the first target node concurrently processing the service, and the ideal visitor information of the user of the target area to the service The number of resource instances that the first target node should allocate for the service.
The method of claim 4, wherein determining the ideal amount of access information of the user of the target area to the service when the service is deployed on the first target node comprises:

Determining, if the service is deployed on the first target node, obtaining the ideal visitor information of the user of the target area according to the access quantity information of the target user in the target area. ;

Wherein, for each user in the target area, if the distance between the location where the user is located and the location of the first target node is within a preset distance range, the user is determined to be the target user.
The method according to claim 4 or 5, wherein the amount of access to the service by the user of the target area, the number of tasks for processing the service by the first target node concurrently, and the target The ideal access information of the user of the area determines the number of resource instances that the first target node should allocate for the service, including:

a minimum value of the user's access amount information for the service, the number of tasks for which the first target node concurrently processes the service, and the user's ideal access amount information for the service Determined as: the number of resource instances that the first target node should allocate for the service.
The method according to any one of claims 1 to 6, wherein after the determined resource instance corresponding to the number of the resource instances on the first target node is allocated to the service, the method further includes:

If the constraint corresponding to the service exists, and the constraint includes: deploying the service to a second target node of the M nodes, where the second target node and the first target node Different, the service is deployed on the second target node.
The method of claim 7, wherein the constraint further comprises: allocating a target number of resource instances to the service on the second target node;

Deploying the service on the second target node includes:

If the number of available resource instances on the second target node is less than the target number, execute:

Deleting a service that has been deployed on the second target node and minimizing the number of allocated resource instances from the second target node until the number of available resource instances on the second target node is greater than the target number ;

Deploying the service on the second target node and allocating the target number of resource instances to the service.
The method according to any one of claims 1 to 8, wherein after the K services to be deployed are deployed to the M nodes, the method further includes:

Obtaining system resources corresponding to the K services according to the number of resource instances allocated by the K services, and the user access amount information corresponding to the K services after the K services to be deployed are deployed to the M nodes Utilization rate

If it is determined that the system resource utilization is less than a preset resource utilization threshold, the K services are redeployed.
A service deployment device in an edge cloud architecture, comprising: a memory and a processor, wherein:

a memory for storing computer program instructions;

a processor, configured to invoke the computer program instructions stored in the memory, to perform the following processing:

When the K services to be deployed are deployed to the M nodes, for each of the K services to be deployed, performing: estimating, according to the historical log information, the users of each of the N regions Service access information; The N is an integer greater than or equal to 1;

Determining, for each of the M nodes, a total communication delay required for a user of the N areas to access the service after the service is deployed in the node according to the access information, M total communication delays;

Determining, from the M nodes, a first target node that the service should be deployed according to the M total communication delays;

Determining, according to the number of available resource instances corresponding to the first target node, and the user access information of the service in each of the N regions, that the first target node should be allocated for the service The number of resource instances;

Allocating the determined resource instance corresponding to the quantity of the resource instances on the first target node to the service;

The K and M are integers greater than or equal to 1.
The service deployment apparatus according to claim 10, wherein the processor determines, after each of the M nodes, that the service is deployed in the node according to the access amount information, The total communication delay required by the users of the N areas to access the service, and the M total communication time delays are obtained, specifically for:

Performing each of the M nodes as the current node, respectively, performs:

Determining, according to the access information and the communication performance indicator of the current node, a basic communication delay of the user of the N areas accessing the service after the service is deployed in the current node;

If it is determined that the node currently deployed by the service is different from the current node, determining a migration communication delay corresponding to a process of migrating the service from the currently deployed node to the current node;

And determining, according to the basic communication delay and the migration communication delay, a total communication delay required for a user of the N areas to access the service.
The service deployment apparatus according to claim 10 or 11, wherein the processor is in accordance with the number of available resource instances corresponding to the first target node, and the user pair of each of the N areas The traffic information of the service, when determining the number of resource instances that the first target node should allocate for the service, specifically for:

Determining, according to the performance indicator of the service, the amount of access to the service by the user of each of the N areas, the resource instance consumption corresponding to the service;

Determining, according to the number of available resource instances corresponding to the first target node, and the resource instance consumption corresponding to the service, the number of tasks that the first target node can concurrently process the service;

Determining that the first target node should be the service according to the number of tasks of the first target node concurrently processing the service, and the user access information of the service of each of the N areas The number of resource instances assigned.
The service deployment apparatus according to claim 12, wherein said processor concurrently processes a number of tasks of said service according to said first target node, and a user pair of each of said N areas The traffic information of the service, when determining the number of resource instances that the first target node should allocate for the service, specifically for:

Determining, as the target area, an area in the N areas that has the largest amount of access information of the service;

Determining an ideal amount of visitor information of the user of the target area to the service if the service is deployed on the first target node;

Determining, according to the information about the access of the service by the user of the target area, the number of tasks of the first target node concurrently processing the service, and the ideal visitor information of the user of the target area to the service The first The number of resource instances that a target node should allocate for the service.
The service deployment apparatus according to claim 13, wherein said processor determines an ideal amount of access to said service by said user of said target area when said service is deployed on said first target node When information is used, it is specifically used to:

Determining, if the service is deployed on the first target node, obtaining the ideal visitor information of the user of the target area according to the access quantity information of the target user in the target area. ;

Wherein, for each user in the target area, if the distance between the location where the user is located and the location of the first target node is within a preset distance range, the user is determined to be the target user.
The service deployment apparatus according to claim 13 or 14, wherein the processor concurrently processes the service according to the access amount information of the user of the target area to the first target node The number of tasks, and the ideal amount of access information of the user of the target area to the service, when determining the number of resource instances that the first target node should allocate for the service, specifically for:

a minimum value of the user's access amount information for the service, the number of tasks for which the first target node concurrently processes the service, and the user's ideal access amount information for the service Determined as: the number of resource instances that the first target node should allocate for the service.
The service deployment apparatus according to any one of claims 10 to 15, wherein the processor allocates the determined resource instance corresponding to the number of resource instances on the first target node to the service Later, it is also used to:

If the constraint corresponding to the service exists, and the constraint includes: deploying the service to a second target node of the M nodes, where the second target node and the first target node Different, the service is deployed on the second target node.
The service deployment apparatus according to claim 16, wherein the constraint further comprises: allocating a target number of resource instances to the service on the second target node;

When the processor deploys the service on the second target node, specifically:

If the number of available resource instances on the second target node is less than the target number, execute:

Deleting a service that has been deployed on the second target node and minimizing the number of allocated resource instances from the second target node until the number of available resource instances on the second target node is greater than the target number ;

Deploying the service on the second target node and allocating the target number of resource instances to the service.
The service deployment device according to any one of claims 10 to 17, wherein after the processor deploys the K services to be deployed to the M nodes, the processor is further configured to:

Obtaining system resources corresponding to the K services according to the number of resource instances allocated by the K services, and the user access amount information corresponding to the K services after the K services to be deployed are deployed to the M nodes Utilization rate

If it is determined that the system resource utilization is less than a preset resource utilization threshold, the K services are redeployed.
A service deployment device in an edge cloud architecture, comprising:

The load analysis module is configured to estimate, for each service of the K services to be deployed, each of the N regions according to the historical log information when the K services to be deployed are deployed to the M nodes. User access information to the service; the N is an integer greater than or equal to 1; for each of the M nodes, determining, according to the access amount information, that the service is deployed after the node Describe the total communication delay required by the users of the N areas to access the service, and obtain M total communication delays;

a deployment decision module, configured to determine the service from the M nodes according to the M total communication delays a first target node to be deployed; determining the first according to the number of available resource instances corresponding to the first target node, and the amount of access information of the user of each of the N regions to the service The number of resource instances that the target node should allocate for the service; and the determined resource instance corresponding to the number of the resource instances on the first target node is allocated to the service;

The K and M are integers greater than or equal to 1.
The service deployment apparatus according to claim 19, wherein the load analysis module determines, after each of the M nodes, that the service is deployed after the node according to the visitor information. The total communication delay required by the users of the N areas to access the service, and the M total communication time delays are obtained, specifically for:

Performing each of the M nodes as the current node, respectively, performs:

Determining, according to the access information and the communication performance indicator of the current node, a basic communication delay of the user of the N areas accessing the service after the service is deployed in the current node;

If it is determined that the node currently deployed by the service is different from the current node, determining a migration communication delay corresponding to a process of migrating the service from the currently deployed node to the current node;

And determining, according to the basic communication delay and the migration communication delay, a total communication delay required for a user of the N areas to access the service.
The service deployment device according to claim 19 or 20, wherein the deployment decision module is in accordance with the number of available resource instances corresponding to the first target node, and the user pair of each of the N regions The traffic information of the service, when determining the number of resource instances that the first target node should allocate for the service, is specifically used to:

Determining, according to the performance indicator of the service, the amount of access to the service by the user of each of the N areas, the resource instance consumption corresponding to the service;

Determining, according to the number of available resource instances corresponding to the first target node, and the resource instance consumption corresponding to the service, the number of tasks that the first target node can concurrently process the service;

Determining that the first target node should be the service according to the number of tasks of the first target node concurrently processing the service, and the user access information of the service of each of the N areas The number of resource instances assigned.
The service deployment apparatus according to claim 21, wherein said deployment decision module concurrently processes the number of tasks of said service according to said first target node, and user pairs of each of said N areas The traffic information of the service, when determining the number of resource instances that the first target node should allocate for the service, is specifically used to:

Determining, as the target area, an area in the N areas that has the largest amount of access information of the service;

Determining an ideal amount of visitor information of the user of the target area to the service if the service is deployed on the first target node;

Determining, according to the information about the access of the service by the user of the target area, the number of tasks of the first target node concurrently processing the service, and the ideal visitor information of the user of the target area to the service The number of resource instances that the first target node should allocate for the service.
The service deployment apparatus according to claim 22, wherein said deployment decision module determines that said user of said target area has an ideal access to said service when said service is deployed on said first target node When the amount information is used, it is specifically used to:

Determining, if the service is deployed on the first target node, according to target user access in the target area The traffic information of the service obtains ideal visitor information of the user in the target area to the service;

Wherein, for each user in the target area, if the distance between the location where the user is located and the location of the first target node is within a preset distance range, the user is determined to be the target user.
The service deployment apparatus according to claim 22 or 23, wherein the deployment decision module concurrently processes the service according to the user's access amount information of the service in the target area and the first target node The number of tasks, and the ideal amount of access information of the user of the target area to the service, when determining the number of resource instances that the first target node should allocate for the service, specifically for:

a minimum value of the user's access amount information for the service, the number of tasks for which the first target node concurrently processes the service, and the user's ideal access amount information for the service Determined as: the number of resource instances that the first target node should allocate for the service.
The service deployment apparatus according to any one of claims 19 to 24, wherein the deployment decision module allocates the determined resource instance corresponding to the number of resource instances on the first target node to the After the service, it is also used to:

If the constraint corresponding to the service exists, and the constraint includes: deploying the service to a second target node of the M nodes, where the second target node and the first target node Different, the service is deployed on the second target node.
The service deployment apparatus according to claim 25, wherein the constraint further comprises: allocating a target number of resource instances to the service on the second target node;

When the deployment decision module deploys the service on the second target node, specifically:

If the number of available resource instances on the second target node is less than the target number, execute:

Deleting a service that has been deployed on the second target node and minimizing the number of allocated resource instances from the second target node until the number of available resource instances on the second target node is greater than the target number ;

Deploying the service on the second target node and allocating the target number of resource instances to the service.
The service deployment device according to any one of claims 19 to 26, wherein the deployment decision module deploys the K services to be deployed to the M nodes, and is further configured to:

Obtaining system resources corresponding to the K services according to the number of resource instances allocated by the K services, and the user access amount information corresponding to the K services after the K services to be deployed are deployed to the M nodes Utilization rate

If it is determined that the system resource utilization is less than a preset resource utilization threshold, the K services are redeployed.