阅读说明：本技术 一种基于容器的cu与mec共平台部署方法 (Container-based CU and MEC common-platform deployment method ) 是由 孙志伟 于 2021-07-02 设计创作，主要内容包括：本发明涉及一种基于容器的CU与MEC共平台部署方法。本发明可以基于kata容器架构,在CU与MEC共平台的云化、虚拟化5G扩展型小基站部署边缘应用,提供安全性保障的同时,对平台资源进行充分利用的情况下,尽可能多的满足边缘应用部署请求,实现效益最大化。提供容器部署请求和节点双分组排序机制,节约部署请求申请时间。(The invention relates to a container-based CU and MEC common platform deployment method. The invention can be based on the kata container architecture, and can meet the edge application deployment request as much as possible under the condition of fully utilizing platform resources while providing security guarantee by deploying the edge application in the 5G extended small base station which is clouded and virtualized and has the same platform of the CU and the MEC, thereby realizing the maximum benefit. And a container deployment request and node double-grouping ordering mechanism is provided, so that the deployment request application time is saved.)

1. A container-based CU and MEC co-platform deployment method is characterized by comprising the following steps:

step 1, arranging a kata container operation environment, configuring a docker container engine to manage container-shim-kata-v 2, and creating conditions for deploying containers which are isolated from each other;

step 2, acquiring available resource information of each node in a cluster database, wherein the available resource information comprises general resources and special resources, and the special resources at least comprise one resource except a CPU (central processing unit), a memory and a hard disk;

step 3, grouping the nodes according to the most special resources contained in each node, and counting the total amount of available general resources of all the nodes in each group in real time;

step 4, sequencing according to the number of the available special resources corresponding to each group contained in the node from low to high;

step 5, when a container deployment instruction is received, acquiring resource demand information of the container according to configuration information contained in the instruction, wherein the resource demand information comprises general resource application information and special resource application information;

step 6, grouping the container deployment requests according to the most special resource requirements in the instructions and sequencing the container deployment requests from high to low according to the total resource requirement contained in the container deployment requests;

step 7, matching node grouping and container deployment request grouping;

step 8, judging whether a first matching node matched with the deployment request exists in the grouping nodes, wherein various available resource values of the first matching node are not smaller than the request value of the resource;

and 9, directly deploying the container if the matching is successful.

2. The vessel-based CU and MEC co-platform deployment method as claimed in claim 1, further comprising the steps of:

step 10, if the matching is unsuccessful, splitting the deployment request according to each resource request value, wherein each deployment sub-request comprises one of a non-split resource request or a split resource split request;

step 11, returning the sub-requests to the group for reordering;

step 12, judging whether a second matching node matched with each deployment sub-request exists in the node group, wherein the value of each residual available resource of the second matching node is not less than the request value of the matched container deployment sub-request for the resource;

step 13, if the matching is successful, directly deploying the sub-request;

and 14, if the matching is unsuccessful, reallocating the node group for the sub-request until the deployment is completed.

Technical Field

The invention relates to the field of wireless communication and terminals, in particular to a container-based CU and MEC common-platform deployment method.

Background

The base station of the 5G extended small base station has a three-level architecture including a Baseband processing Unit (BBU), a switch (HUB), and a Radio frequency processing Unit (RRU), as shown in fig. 1.

The 5G extended small base station can adopt the cloud and virtualization technology to realize pooling sharing of wireless network resources, can support platform-shared deployment with MEC/UPF (User Plane Function), meets flexible requirements of various 5G indoor services, meets elastic expansion requirements of various services on wireless bottom layer basic resources, and realizes pooling gain.

Meanwhile, the virtualized small base station can well realize software and hardware decoupling, so that 5G multi-class service scenes do not depend on bottom hardware drive and are deployed rapidly. For both of the following baseband implementations, a co-platform deployment of BBUs (CU & DUs) and MECs may be considered. As shown in fig. 2.

The MEC and the BBU are deployed on the same platform, so that the cost for independently deploying the BBU/MEC can be reduced, a hardware resource pool of the system is fully utilized, the resource utilization rate is maximized, the sharing between the data information of the wireless access network and the MEC platform can be realized, and the MEC platform can be better utilized to optimize the resource allocation and intelligent management of the access network system.

Meanwhile, the MEC and BBU common platform deployment can well meet the requirements of the network and the commercial complex in the aspects of service innovation, operation innovation and upgrading by utilizing the advantages of sinking content and application, edge network service processing capacity, low-delay experience, high-precision indoor positioning and the like.

For large-scale cluster edge application deployment of the MEC platform, a traditional virtual machine occupies a large amount of software and hardware resources, and the resource utilization rate is reduced, so that a light-weight virtualization technology is particularly important.

The container technology is a portable virtualization technology for lightweight packaging application, saves more computing resources than a virtual machine technology, and is more flexible. Docker is used as an open source container engine, and based on a kernel lightweight virtualization technology, application resource isolation and configuration can be achieved. The Docker container architecture is shown in fig. 3.

Meanwhile, the 5G indoor coverage oriented access network has open capability, needs to meet the localization requirement of indoor vertical industry application, and solves the problems of localization intensive communication and local data privacy. To address the data privacy issue, Kata containers are a good quality solution.

Kata containers provide inter-container isolation by using hardware virtualization. In the case of the Docker engine, container-shim-kata-v 2 provides virtual machine isolation at the container level. Each container is started as a lightweight virtual machine and has a unique kernel. Since each container is now running with its own virtual machine, they are not able to access the host kernel and can gain all the security advantages of the virtual machine.

As shown in the architecture diagram of the kata container in FIG. 4, because the kata container has strong independence and the hardware is isolated from each other, the resource utilization rate of different containers running on different nodes (single computers) is different. The resource utilization rate of each container determines the resource utilization efficiency of the whole cluster, so that cluster resource scheduling is necessary. When a container with multiple resource requirements is started, if the node resource cannot meet the requirement of the container, the container cannot be operated. And resources of other dimensions of the nodes are left to form fragment resources, so that the fragment resources cannot be effectively utilized, and waste is caused.

Disclosure of Invention

The invention aims to provide a container-based CU and MEC common-platform deployment method aiming at the requirements of optimizing resource utilization rate and protecting data privacy when edge application common-platform deployment is carried out on a virtual substation based on a common server on CU and MEC.

In order to achieve the purpose, the technical scheme of the invention is as follows: a container-based CU and MEC co-platform deployment method, comprising the steps of:

step 1, arranging a kata container operation environment, configuring a docker container engine to manage container-shim-kata-v 2, and creating conditions for deploying containers which are isolated from each other;

step 2, acquiring available resource information of each node in a cluster database, wherein the available resource information comprises general resources and special resources, and the special resources at least comprise one resource except a CPU (central processing unit), a memory and a hard disk;

step 3, grouping the nodes according to the most special resources contained in each node, and counting the total amount of available general resources of all the nodes in each group in real time;

step 4, sequencing according to the number of the available special resources corresponding to each group contained in the node from low to high;

step 5, when a container deployment instruction is received, acquiring resource demand information of the container according to configuration information contained in the instruction, wherein the resource demand information comprises general resource application information and special resource application information;

step 6, grouping the container deployment requests according to the most special resource requirements in the instructions and sequencing the container deployment requests from high to low according to the total resource requirement contained in the container deployment requests;

step 7, matching node grouping and container deployment request grouping;

step 8, judging whether a first matching node matched with the deployment request exists in the grouping nodes, wherein various available resource values of the first matching node are not smaller than the request value of the resource;

and 9, directly deploying the container if the matching is successful.

In an embodiment of the present invention, the method further includes the following steps:

step 10, if the matching is unsuccessful, splitting the deployment request according to each resource request value, wherein each deployment sub-request comprises one of a non-split resource request or a split resource split request;

step 11, returning the sub-requests to the group for reordering;

step 12, judging whether a second matching node matched with each deployment sub-request exists in the node group, wherein the value of each residual available resource of the second matching node is not less than the request value of the matched container deployment sub-request for the resource;

step 13, if the matching is successful, directly deploying the sub-request;

and 14, if the matching is unsuccessful, reallocating the node group for the sub-request until the deployment is completed.

Compared with the prior art, the invention has the following beneficial effects:

1. based on the kata container operation environment, the containers in the nodes are prevented from being accessed with each other, and data privacy is protected.

2. Node resources are fully utilized, and node resource waste is avoided.

3. And reasonably planning cluster resources, deploying to different nodes according to container requirements, splitting when necessary, and avoiding cluster resource waste.

4. A node & request dual-grouping ordering mechanism is introduced, so that the time of deployment request is saved, and resource fragmentation caused by occupation of high-resource nodes by low-demand deployment requests is avoided.

Drawings

FIG. 1 is a 3-level architecture diagram of a 5G expansion type small station of a room;

FIG. 2 is a schematic diagram of a unified computing environment for an indoor scenario;

FIG. 3 is a diagram of a Docker-run container architecture;

FIG. 4 is a kata container architecture diagram;

FIG. 5 is a schematic flow chart of a container deployment method of the present invention 1;

FIG. 6 is a schematic flow chart of a container deployment method of the present invention, FIG. 2;

FIG. 7 is a schematic diagram of an embodiment of the present invention.

Detailed Description

The technical scheme of the invention is specifically explained below with reference to the accompanying drawings.

The invention relates to a container-based CU and MEC common-platform deployment method, which comprises the following steps:

step 1, arranging a kata container operation environment, configuring a docker container engine to manage container-shim-kata-v 2, and creating conditions for deploying containers isolated from each other.

And 2, acquiring available resource information of each node in a cluster database, wherein the available resource information comprises general resources and special resources, and the special resources at least comprise one resource except a CPU (Central processing Unit), a memory and a hard disk.

And 3, grouping the nodes according to the most special resources contained in each node, and counting the total amount of available general resources of all the nodes in each group in real time.

And 4, sequencing according to the number of the available special resources corresponding to each group contained in the node from low to high.

And 5, when a container deployment instruction is received, acquiring resource demand information of the container according to configuration information contained in the instruction, wherein the resource demand information comprises general resource application information and special resource application information.

And 6, grouping the container deployment requests according to the most special resource requirements in the instructions and sequencing the container deployment requests from high to low according to the total resource requirement contained in the container deployment requests.

And 7, matching the node grouping and the container deployment request grouping.

And 8, judging whether a first matching node matched with the deployment request exists in the grouping nodes, namely the first matching node, wherein all available resource values of the first matching node are not smaller than the request value of the resource.

And 9, directly deploying the container if the matching is successful.

As shown in fig. 5.

And step 10, if the matching is unsuccessful, splitting the deployment request according to each resource request value, wherein each deployment sub-request comprises one of a non-split resource request or a split resource split request.

And step 11, returning the sub-requests to the group for reordering.

And step 12, judging whether a second matching node matched with each deployment sub-request exists in the node group, wherein the value of each residual available resource of the second matching node is not less than the request value of the matched container deployment sub-request for the resource.

And step 13, if the matching is successful, directly deploying the sub-request.

And 14, if the matching is unsuccessful, reallocating the node group for the sub-request until the deployment is completed.

As shown in fig. 6.

The following is a specific implementation of the present invention.

The embodiment of the invention is an embodiment of deploying a container before deploying edge application on a general server cluster of a common platform of an MEC and a BBU (CU & DU). The schematic diagram is shown in fig. 7, and the main ideas are as follows:

step 1, configuring kata container operation environment to ensure that containers in each node cannot access each other,

and 2, the cluster node grouping module acquires the current available resource information of each node in the cluster from a preset database, wherein the current available resource information comprises general resources and special resources, and then groups each computing node according to the maximum one of the available special resources of each node, so that the available general resource information of all the nodes in each group is counted in real time.

And 3, the cluster node sequencing module sequences the nodes in the group from low to high according to the available special resources of each node.

And 4, the container request grouping module receives the deployment request of the container and the sub-request for quitting grouping, acquires resource demand information of the deployment container according to configuration information included in the container deployment request, wherein the resource demand information includes general resource demand information and special resource demand information, and groups the container deployment instructions according to the most special resource requests in the container requests.

And 5, the container request ordering module orders the container request groups according to the total amount of resources requested by the container deployment requests/sub-requests from high to low.

And 6, matching the container deployment request group and the node group according to the special resource information by the grouping matching module.

And 7, the node-request matching module deploys the resource requests to sequentially send resource application information to each computing node in the node group according to the respective sequence of the requests and the nodes, and the nodes request the resource from high to low and match the owned resource from low to high according to the sequence of the instruction.

And 8, when receiving the information of successful matching returned by the nodes in the node group, the container deployment module determines the node returning the information of successful matching as a first matching node, wherein the resource residual value of each resource of the first matching node is not less than the resource request value of the resource. A container is deployed in the first matching node.

And 9, if the container deployment module does not receive the information of successful matching, indicating that no first matching node exists. And sending the container deployment request to a container request splitting module, splitting the container deployment request into a plurality of deployment sub-requests according to the detachable value of each resource, wherein each deployment sub-request comprises a resource sub-request value of each resource, and the resource sub-request value of each resource comprises the non-detachable value of the resource request value of the resource or one of a plurality of split request values split by the detachable value.

And step 10, the container request ordering module continues to queue the matching nodes in the request group according to the resource requirements of the sub-requests.

Step 11, when receiving the information of successful matching returned by the nodes in the node group, determining the node returning the information of successful matching as a second matching node,

and step 12, the container deployment module deploys the container sub-request in the second matching node.

Step 13, after each request/sub-request deployment is completed, the cluster node sequencing module recalculates the available special resources of the nodes, and if the special resources corresponding to the groups are remained, the cluster node sequencing module performs sequencing again in the groups; and if the special resources corresponding to the grouping are exhausted, the nodes are quitted from the grouping and returned to the cluster node grouping module.

And step 14, if the matching is finished and the message of successful matching is not received, returning the sub-request to the container request grouping module for grouping again, and queuing and matching as the sub-request in another request grouping. Until deployment is complete.

The above are preferred embodiments of the present invention, and all changes made according to the technical scheme of the present invention that produce functional effects do not exceed the scope of the technical scheme of the present invention belong to the protection scope of the present invention.