A method and device for implementing cluster services in container network

By generating virtual IP and POD groups in the SDN controller, and using the smart network card OVS flow table to forward cluster service traffic, the problems of host memory consumption and low performance caused by iptables are solved, and efficient traffic forwarding and security protection are achieved.

CN116233014BActive Publication Date: 2024-11-26CHINA TELECOM CLOUD TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310225860.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-03-09
Publication Date
2024-11-26
Estimated Expiration
2043-03-09

AI Technical Summary

Technical Problem

In the prior art, using iptables to process the traffic of cluster services has resulted in excessive consumption of host memory resources and low performance, which cannot effectively solve the problems of traffic forwarding efficiency and security protection.

Method used

By generating virtual IP in the SDN controller, forming a POD group, and using the OVS stream table on the smart network card to forward traffic, issuing security group and QOS functions, traffic release and speed limit are achieved, and host resource consumption is avoided.

Benefits of technology

It effectively solves the problem of host memory resource consumption caused by insufficient iptables performance, improves traffic forwarding efficiency, and provides security protection and speed limiting functions for cluster service traffic.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116233014B_ABST
    Figure CN116233014B_ABST
Patent Text Reader

Abstract

The present invention discloses a method and device for implementing cluster services in a container network, and relates to the field of network address translation in data communication. The method for implementing cluster services in a container network comprises the following steps: generating a corresponding virtual IP in an SDN controller according to the IP address of a cluster service created by a user; forming a group with a plurality of PODs corresponding to the cluster service backend that send VRRP messages through the SDN controller; publishing routing information of the group's virtual IP externally on the gateway VPP; sending the traffic to the controller when there is traffic accessing the virtual IP address, and the controller performs flow-by-flow load sharing to a plurality of PODs in the group according to the traffic; issuing relevant flow tables to the corresponding OVS to implement traffic forwarding; issuing security group functions and QOS functions to the virtual IP through the SDN controller to implement management of traffic release and speed limit of the virtual IP. The present invention realizes that host memory resources are not consumed.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of network address translation in data communication, and in particular to a method and device for implementing cluster services in a container network. Background Art

[0002] Tianyi Cloud's new generation IaaS cloud platform CloudOS4.0 supports multi-AZ resource layout, ultra-large cluster scale (single AZ supports more than 10,000 units) and software and hardware collaboration. The core capabilities of the resource pool meet the requirements of ultra-large-scale, high-performance, secure and reliable cloud services. It improves the management and scheduling capabilities of the ultra-large-scale resource pool of the IaaS platform, supports ultra-large-scale data center networking and multi-AZ architecture, has the network management capabilities of a resource pool with a scale of 5,000 nodes, supports a scale of 5,000 servers in a single AZ, and increases the Region scale by more than 10 times, reaching the advanced level in the industry. It meets the business needs of enterprises for multi-active and disaster recovery in two locations and three centers. CloudOS4.0, a new generation cloud platform developed based on the national cloud, is fully commercialized and deployed on the current network. As an autonomous and controllable cloud base of "one cloud with multiple states and one cloud with multiple cores", as one of the core components of Tianyi Cloud 4.0, it provides a high-performance, high-availability and high-reliability basic cloud network base. Containerized applications will deploy multiple services on the same node, and each service requires its own network space. To avoid conflicts with other business networks, the Pod needs to have its own independent network space. If the applications in the Pod need to communicate with other networks, the Pod needs to be able to access different networks.

[0003] The existing method for processing traffic is generally to use the ipvlan l2 mode. The ENI plug-in creates an ipvlan sub-interface on the ENI and places the sub-interface in each Pod, while setting the EIP, routing, and gateway information in the Pod. In particular, in the host namespace, an ipvlan sub-interface is reserved for processing service traffic. Since the network of the sub-interface in the Pod is consistent with the network of the host ENI, the inbound and outbound traffic of the Pod is forwarded from the sub-interface to the ENI main interface, and is forwarded completely according to the normal network of the ENI. For traffic between pods on the same node: the traffic does not leave the node, but is directly forwarded between different sub-interfaces through the main interface ENI of this node; for traffic between pods across nodes: the egress traffic of the Pod on node 1 is first forwarded to the main interface ENI A through the sub-interface. Since the main interface is the physical network interface of the virtual machine, this part of the traffic is processed completely according to the normal traffic of the virtual machine. The traffic is processed by OVS and finally forwarded to the main interface ENI B of node 2. The host port automatically forwards the traffic to the corresponding sub-interface, and the traffic is successfully forwarded to the corresponding Pod. Among them, Kubernets needs to process the Service traffic through kube-proxy (iptables). The scope of kube-proxy is the host protocol stack of the host machine, that is, the host network namespace. Therefore, in order to reuse this part of the processing flow, the Service traffic needs to be forwarded to the host network namespace so that kube-proxy can process it as expected. In summary, the current solution is to create a special ipvlan subinterface for ENI, which is retained in the host network namespace. At the same time, TC rules are added on the main interface to identify the Service traffic and forward this part of the traffic to this subinterface.

[0004] However, the inventors of this application found that the above technology has at least the following technical problems in the process of implementing the technical solution of the invention in the embodiment of this application:

[0005] In the prior art, cluster service uses iptables to process traffic. However, due to the general performance of iptables itself, a large amount of host memory resources are consumed. Summary of the invention

[0006] The embodiments of the present application provide a method and device for realizing cluster services in a container network, thereby solving the problem in the prior art that a large amount of host memory resources are consumed due to the general performance of iptables itself, and realizing cluster service traffic forwarding through the OVS flow table on the smart network card without consuming resources on the host.

[0007] The present application provides a method for implementing a cluster service in a container network, including the following steps:

[0008] S1, generates a corresponding virtual IP in the SDN controller according to the IP address of the cluster service created by the user;

[0009] S2, forming a group of multiple PODs corresponding to the cluster service backend that send VRRP packets through the SDN controller;

[0010] S3, publish the routing information of the group's virtual IP on the gateway VPP;

[0011] S4: When traffic accesses the virtual IP address, the traffic is sent to the controller, which then load-balances the traffic to multiple PODs in the group.

[0012] S5, sends the relevant flow table to the corresponding OVS to implement traffic forwarding;

[0013] S6, through the SDN controller to send security group function to the virtual IP, to achieve the management of traffic release;

[0014] S7, the SDN controller sends the QOS function to the virtual IP and limits the speed of the virtual IP.

[0015] Furthermore, the IP address of the cluster service in S1 is the same as the virtual IP address of the SDN controller.

[0016] Furthermore, the POD in S3 exists in a virtual machine, and the virtual machine includes a virtual network card for communication between PODs.

[0017] Furthermore, the OVS in the S5 exists on the smart network card and is used to forward the traffic of the cluster service.

[0018] Furthermore, when the security group function is issued to the virtual IP in S6, a flow table corresponding to the security group is generated.

[0019] Furthermore, the flow table corresponding to the security group includes an outbound security group and an inbound security group.

[0020] Furthermore, the management of traffic release in S6 refers to releasing only the security group corresponding to the virtual IP of the cluster service, and not releasing other traffic.

[0021] Furthermore, the virtual IP address in S7 corresponds to a meter table for limiting the bandwidth under UDP and TCP by bps, and the meter table is set in x86.

[0022] Furthermore, the speed limit standard for the virtual IP in S7 is 500000 kbps, and data packets with a bps exceeding 500000 kbps will be discarded.

[0023] The embodiment of the present application provides a device for implementing cluster services in a container network, including a generation module, a grouping module, a publishing module, a traffic diversion module, a forwarding module, a release module, and a speed limiting module:

[0024] Generation module: used to generate the corresponding virtual IP in the SDN controller according to the IP address of the cluster service created by the user;

[0025] Group building module: used to form a group with multiple PODs that send VRRP packets through the SDN controller in the cluster service backend;

[0026] Publishing module: used to publish the routing information of the virtual IP of this group on the gateway VPP;

[0027] Traffic distribution module: When traffic accesses the virtual IP address, it sends the traffic to the controller, which then distributes the traffic to multiple PODs in the group based on the traffic flow.

[0028] Forwarding module: used to send relevant flow tables to the corresponding OVS to achieve traffic forwarding;

[0029] Release module: used to send security group functions to virtual IP through SDN controller to manage traffic release;

[0030] Speed ​​limit module: used to send QOS function to virtual IP through SDN controller and limit the speed of virtual IP.

[0031] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0032] 1. Since the intelligent network card is used to replace the iptables in the traditional solution to forward traffic, the consumption of host memory resources is avoided, and the problem of large-scale consumption of host memory resources due to the general performance of iptables itself in the prior art is effectively solved, thereby achieving no consumption of host memory resources.

[0033] 2. Since the security group function and QOS function are issued to the virtual IP through the SDN controller, the security group and QOS functions can be provided for the virtual IP of the clusterservice traffic, which effectively solves the problem of insufficient security protection for forwarding traffic in the prior art, and thus realizes security protection for the cluster service traffic.

[0034] 3. Since the OVS flow table on the smart network card is used to forward traffic, traffic forwarding is more direct, which effectively solves the problem of low traffic forwarding efficiency in the existing technology, thereby improving forwarding efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A flow chart of a method for implementing cluster services in a container network provided in Embodiment 1 of the present application;

[0036] Figure 2 This is an overall flow chart of the traditional access service provided in Example 1 of the present application;

[0037] Figure 3 This is a diagram of the internal architecture of the ENI plug-in in the traditional solution provided in Example 1 of the present application;

[0038] Figure 4 A flow chart of traffic forwarding based on a smart network card provided in Example 1 of the present application;

[0039] Figure 5 A diagram of the smart network card forwarding traffic framework provided in Example 1 of the present application;

[0040] Figure 6 The forwarding flow table diagram corresponding to host1 provided in Example 1 of the present application;

[0041] Figure 7 The forwarding flow table diagram corresponding to host2 provided in Example 1 of the present application;

[0042] Figure 8 The forwarding flow table diagram corresponding to host3 provided in Example 1 of the present application;

[0043] Fig. 9 The flow table diagram corresponding to the security group provided in Example 1 of the present application;

[0044] Fig.10 This is a structural diagram of a device for implementing cluster services in a container network provided in Example 1 of the present application. DETAILED DESCRIPTION

[0045] The embodiments of the present application provide a method and device for implementing cluster services in a container network, thereby solving the problem in the prior art that a large amount of host memory resources are consumed due to the general performance of iptables itself. The flow table of OVS is used on the smart network card to implement cluster service traffic forwarding without consuming the host's memory resources.

[0046] The technical solution in the embodiment of the present application is to solve the problem that a large amount of host memory resources are consumed due to the general performance of iptables itself. The overall idea is as follows:

[0047] The corresponding virtual IP is generated in the SDN controller according to the IP address of the cluster service created by the user; the multiple POD containers corresponding to the cluster service backend send VRRP packets through the controller to form a group; then the routing information of the group's virtual IP is published on the gateway VPP; when there is traffic accessing the virtual IP address, the traffic is sent to the controller, and the controller performs flow-by-flow load sharing to multiple PODs in the group according to the traffic; then the relevant flow table is sent to the corresponding OVS to realize traffic forwarding; the SDN controller sends the security group function to the virtual IP to realize the management of traffic release and limit the speed of the virtual IP, thereby avoiding the consumption of host memory resources.

[0048] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0049] Embodiment 1

[0050] like Figure 1 As shown, it is a flow chart of a method for implementing a cluster service in a container network provided by an embodiment of the present application. The method is applied to a device for implementing a cluster service in a container network. The method includes the following steps:

[0051] S1, generates a corresponding virtual IP in the SDN controller according to the IP address of the cluster service created by the user;

[0052] S2, forming a group of multiple PODs corresponding to the cluster service backend that send VRRP packets through the SDN controller;

[0053] S3, publish the routing information of the group's virtual IP on the gateway VPP;

[0054] S4: When traffic accesses the virtual IP address, the traffic is sent to the controller, which then load-balances the traffic to multiple PODs in the group.

[0055] S5, sends the relevant flow table to the corresponding OVS to implement traffic forwarding;

[0056] S6, through the SDN controller to send security group function to the virtual IP, to achieve the management of traffic release;

[0057] S7, the SDN controller sends the QOS function to the virtual IP and limits the speed of the virtual IP.

[0058] In this embodiment, SDN refers to Software Defined Network, clusterservice refers to cluster service, and OVS refers to Open Vswitch, that is, virtual switch;

[0059] like Figure 2 As shown in the overall flow chart of traditional access to Service, the overall processing flow includes K8s cluster, namely Kubernets, and Kubernets needs to process Service traffic through kube-proxy (iptables). Among them, the Service network supports the following multiple modes to access clients of different sources and types respectively: 1) ClusterIP type Service is used for access between applications within the cluster. If your application needs to be exposed to the cluster to provide services, you need to use ClusterIP type Service exposure, which means that the default Service type when creating Service is ClusterIP; 2) NodePort type Service exposes the application deployed in the cluster to the outside through a fixed port on the cluster node, so that it can be accessed outside the cluster through the node IP and this fixed port; 3) LoadBalancer type Service also exposes the application deployed in the cluster to the outside, but it is exposed through Alibaba Cloud's load balancing, which has higher availability and performance than the NodePort method; 4) Headless Service type Service specifies the clusterIP field as None type in the Service attribute, using Headless After the Service type is set, the Service will not have a fixed virtual IP address. When the client accesses the Service domain name, the IP addresses of all backend Pod instances will be returned through DNS. The client needs to use DNS load balancing to achieve backend load balancing; 5) ExternalName type Service maps the domain name outside the cluster to the Service inside the cluster. For example, the external database domain name is mapped to the Service name inside the cluster, so that it can be directly accessed through the Service name inside the cluster;

[0060] Since containerized applications deploy multiple services on the same node, and each service requires its own network space, in order to avoid conflicts with other service networks, POD needs to have its own independent network space. Applications in Pod need to communicate with other networks, so POD needs to be able to access different networks. Figure 3 As shown in the internal architecture diagram of the ENI plug-in in the traditional solution, the traditional solution ENI plug-in adopts the ipvlan 12 mode, the POD is located in the virtual machine, and the virtual machine is located in the host.

[0061] Furthermore, the IP address of the cluster service in S1 is the same as the virtual IP address of the SDN controller.

[0062] In this embodiment, the IP address of the cluster service created by the user is 10.1.1.252, which exists as a virtual IP on the SDN controller, that is, the virtual IP is also 10.1.1.252. Figure 4 As shown in the flow chart of traffic forwarding based on smart network cards: the containers corresponding to the backend of cluster service are POD11, POD13 and POD14. The three PODs send VRRP packets through the controller to form a group, and publish the routing information of the virtual IP 10.1.1.252 of this group on the gateway VPP. When there is traffic accessing the address 10.1.1.252, the traffic is first sent to the controller. The controller performs flow-by-flow load sharing to POD11, POD13 and POD14 according to the traffic, and sends the relevant flow table to the corresponding OVS, thereby realizing traffic forwarding.

[0063] Furthermore, the PODs in S3 exist in a virtual machine, and the virtual machine includes a virtual network card for communication between PODs.

[0064] In this embodiment, if Figure 5 As shown in the smart network card forwarding traffic framework diagram, the container pod is created in the virtual machine, and the address of each pod communicates with other pods through a virtual network card of the virtual machine.

[0065] Furthermore, the OVS in S5 exists on the smart network card and is used to forward the traffic of the cluster service.

[0066] In this embodiment, the OVS flow table on the smart network card is used to forward the traffic of the cluster service, where the forwarding flow tables corresponding to the three hosts are as follows: Figure 6 , 7 and 8.

[0067] Furthermore, when the security group function is issued to the virtual IP in S6, a flow table corresponding to the security group is generated.

[0068] The flow table corresponding to the security group includes the outbound security group and the inbound security group.

[0069] Furthermore, the management of traffic release in S6 means that only the security group corresponding to the virtual IP of the cluster service is released, and other traffic is not released.

[0070] In this embodiment, the SDN controller can issue security group functions to the virtual IP 10.1.1.252 of the cluster service to allow traffic from the security group 10.1.1.252, but not from other security groups. The specific security group corresponding flow table is as follows: Fig. 9 shown.

[0071] Furthermore, the virtual IP address in S7 corresponds to a meter table to limit the bandwidth under UDP and TCP by bps, and the meter table is set in x86.

[0072] Furthermore, the speed limit standard for the virtual IP in S7 is 500000 kbps, and data packets with bps exceeding 500000 kbps will be discarded.

[0073] In this embodiment, the SDN controller issues the QOS function to the virtual IP 10.1.1.252 of the cluster service, and limits the speed of the virtual IP to 500,000 kbps. Since one IP address corresponds to one meter table, in order to limit the bandwidth under UDP and TCP by bps, it is necessary to set the meter table in x86. Here, meter table 1 is set to discard all data packets with bps exceeding 500,000 kbps. To implement this operation, the following code can be used: ovs-ofctl add-meterbr-int meter=1,kbps,band=type=drop,rate=500000-O OpenFlow13.

[0074] like Fig.10 As shown, it is a structural diagram of a device for implementing a cluster service in a container network provided by an embodiment of the present application. The device for implementing a cluster service in a container network provided by an embodiment of the present application includes a generation module, a grouping module, a publishing module, a diversion module, a forwarding module, a release module and a speed limit module:

[0075] Generation module: used to generate the corresponding virtual IP in the SDN controller according to the IP address of the cluster service created by the user;

[0076] Group building module: used to form a group with multiple PODs that send VRRP packets through the SDN controller in the cluster service backend;

[0077] Publishing module: used to publish the routing information of the virtual IP of this group on the gateway VPP;

[0078] Traffic distribution module: When traffic accesses the virtual IP address, it sends the traffic to the controller, which then distributes the traffic to multiple PODs in the group based on the traffic flow.

[0079] Forwarding module: used to send relevant flow tables to the corresponding OVS to achieve traffic forwarding;

[0080] Release module: used to send security group functions to virtual IP through SDN controller to manage traffic release;

[0081] Speed ​​limit module: used to send QOS function to virtual IP through SDN controller and limit the speed of virtual IP.

[0082] In this embodiment, first, the generation module generates a corresponding virtual IP in the SDN controller according to the IP address of the clusterservice created by the user; then, the grouping module forms a group with the corresponding multiple PODs that send VRRP messages through the SDN controller in the cluster service backend; then, the publishing module publishes the routing information of the virtual IP of this group on the gateway VPP; then, when there is traffic accessing the virtual IP address, the diversion module sends the traffic to the controller, and the controller performs flow-by-flow load sharing to multiple PODs in the group according to the traffic; then, the forwarding module sends the relevant flow table to the corresponding OVS to realize traffic forwarding; in the release module, the SDN controller sends the security group function to the virtual IP to realize the management of traffic release; in the speed limit module, the SDN controller sends the QOS function to the virtual IP to limit the speed of the virtual IP.

[0083] The technical solution in the above-mentioned embodiment of the present application has at least the following technical effects or advantages: by utilizing the flow table method of OVS on the smart network card to realize the traffic forwarding of the cluster service, the function of providing security group and QOS for the virtual IP of the cluster service traffic is realized, while also avoiding the problem of consuming a large amount of host memory resources caused by using iptables in the traditional method. In addition, using the smart network card to forward the traffic can also improve the forwarding efficiency.

[0084] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0086] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0087] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0088] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0089] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for implementing cluster services in a container network, characterized in that: The following steps are involved: S1, generates a corresponding virtual IP in the SDN controller according to the IP address of the cluster service created by the user; S2, forming a group of multiple PODs corresponding to the cluster service backend that send VRRP packets through the SDN controller; S3, publish the routing information of the group's virtual IP on the gateway VPP; S4: When traffic accesses the virtual IP address, the traffic is sent to the controller, which then load-balances the traffic to multiple PODs in the group. S5, sends the relevant flow table to the corresponding OVS to implement traffic forwarding; S6, through the SDN controller to send security group function to the virtual IP, to achieve the management of traffic release; S7, the SDN controller sends the QOS function to the virtual IP and limits the speed of the virtual IP; The IP address of the cluster service in S1 is the same as the virtual IP address of the SDN controller; The OVS in S5 exists on the smart network card and is used to forward the traffic of the cluster service.

2. The method for implementing cluster services in a container network as claimed in claim 1, characterized in that: The POD in S3 exists in a virtual machine, and the virtual machine includes a virtual network card for communication between PODs.

3. The method for implementing cluster services in a container network as claimed in claim 1, characterized in that: When the security group function is issued to the virtual IP in S6, a flow table corresponding to the security group is generated.

4. The method for implementing cluster services in a container network as claimed in claim 3, characterized in that: The security group corresponding flow table includes an outbound security group and an inbound security group.

5. The method for implementing cluster services in a container network as claimed in claim 1, characterized in that: The management of traffic release in S6 refers to releasing only the security group corresponding to the virtual IP of the cluster service, and not releasing other traffic.

6. The method for implementing cluster services in a container network as claimed in claim 1, characterized in that: The virtual IP address in S7 corresponds to a meter table that limits the bandwidth under UDP and TCP by bps, and the meter table is set in x86.

7. The method for implementing cluster services in a container network as claimed in claim 1, characterized in that: The standard for limiting the speed of the virtual IP in S7 is 500000 kbps, and data packets with bps exceeding 500000 kbps will be discarded.

8. A device for implementing cluster services in a container network, characterized in that: It includes generation module, group building module, release module, diversion module, forwarding module, release module and speed limit module: Generation module: used to generate the corresponding virtual IP in the SDN controller according to the IP address of the cluster service created by the user; Group building module: used to form a group with multiple PODs that send VRRP packets through the SDN controller in the cluster service backend; Publishing module: used to publish the routing information of the virtual IP of this group on the gateway VPP; Traffic distribution module: When traffic accesses the virtual IP address, it sends the traffic to the controller, which then distributes the traffic to multiple PODs in the group based on the traffic flow. Forwarding module: used to send relevant flow tables to the corresponding OVS to achieve traffic forwarding; Release module: used to send security group functions to virtual IP through SDN controller to manage traffic release; Speed ​​limit module: used to send QOS function to virtual IP through SDN controller and limit the speed of virtual IP; Wherein, the IP address of the cluster service in the generation module is the same as the virtual IP address of the SDN controller; The OVS in the forwarding module exists on the smart network card and is used to forward the traffic of the cluster service.