Method for achieving load balancing of server cluster

Abstract

The invention discloses a method for achieving load balancing of a server cluster. The method for achieving the load balancing of the server cluster includes the following steps that the client broadcasts address resolution protocol request to the server cluster, a load balancing server identifies the address resolution protocol request, and the load balancing server feeds back an physical address to the client when the load balancing server is matched with the address resolution protocol request; the client transmits a service request to the load balancing server according to the physical address of the load balancing server; the load balancing server dispatches the service request to a node server to be processed, wherein the node server is connected with the load balancing server and load value of the node server is the smallest; and after the node server processes the service request, the node server transmits an obtained data package to the client. The method for achieving the load balancing of the server cluster achieves effects of rational utilization of resources and system performance optimization of the server cluster.

Description

technical field [0001] The invention relates to the technical field of computer networks, in particular to a method for realizing server cluster load balancing. Background technique [0002] Cluster, the English name is Cluster. In the field of communication network technology, a cluster refers to a server cluster, which connects multiple servers together to provide the same service, so that the multiple servers in the cluster can work like a single machine from the perspective of the client. Each server in a server cluster is also called a node server. [0003] The server cluster can improve the stability of the system and the data processing capability and service capability of the network center. For example, a server cluster can use multiple servers to perform parallel computing to obtain high computing speed; it can also use multiple computers for backup to avoid single points of failure. In the current server cluster system, the service quality of each node server t...

AI Technical Summary

Problems solved by technology

For the same domain name, different clients will get different addresses, so they also access the Web (network) servers on different addresses, so as to achieve the purpose of load balancing? However, in order to ensure the timely update of DNS data in this method, DNS The refresh time is set to be small, which will cause a large amount of additional network traffic, and the updated DNS data will not take effect immediately; moreover, this method cannot know the differences between the servers of each node, and cannot be used for nodes with better performance. The server allocates more requests, and cannot understand the current status of each node server

[0008] A reverse proxy load balancing method is also provided in the prior art, which uses a proxy server to forward the request to the internal Web server of the cluster, and allows the proxy server to forward the request evenly to one of the multiple internal Web servers, thereby achieving The purpose of load balancing? The disadvantage of reverse proxy load balancing is that for each service request, the proxy server must open two connections (one external connection and one internal connection). When the number of service requests is very large, the load on the proxy server The value increases sharply, and the proxy server becomes the bottleneck of server cluster load balancing

[0009] In addition, the existing technology is based on Network Address Translation (NAT for short) technology, which converts the internal address and external address, so that the server with the internal address can access the external network; and when the server in the external network accesses the address When converting the gateway, the address translation gateway can map its external address to the internal address of an internal server, so as to achieve the purpose of load sharing? Due to the bandwidth sharing of NAT-based load balancing technology, usually the central load balancer has a bandwidth limit, which also becomes Bottleneck for load balancing of the entire server cluster

For example, under the condition of 100MB (MByte, megabyte) Fast Ethernet, this method can reach a bandwidth of 80MB at the fastest; however, in practical applications, this method usually only has an available bandwidth of 40MB to 60MB

Images