A Distributed Service Cluster and Its Routing Synchronization Method

Abstract

The invention discloses a distributed service cluster and a route synchronization method thereof, and relates to the technical field of computers. The distributed service cluster of the present invention includes no less than one server node, and each server node deploys a registration center and no less than one remote procedure call service system. The registration centers of all server nodes form a registration center cluster, and the registration center cluster is used for The global routing table of the remote procedure call service is maintained. The remote procedure call service system is used to maintain the local routing table of the remote procedure call service. The server node provides services according to the global routing table and the local routing table. The distributed service cluster can reduce the complexity of routing synchronization, save operation and maintenance costs and hardware investment.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a distributed service cluster and a route synchronization method thereof. Background technique [0002] For some large Internet companies, the amount of data concurrency is very high, which has already exceeded the limit of the stand-alone service capability. Distributed computing technology is widely used at present, such as distributed storage, cache, message, scheduling, service, etc. Among them, distributed service is a very important distributed computing technology, and it is also a necessary basic technology for major Internet companies. Currently, many distributed service frameworks are used, such as: Dubbo and HSF (short for High-Speed ​​Service Framework), open source high-performance and excellent service frameworks, distributed service framework JSF (short for Java Service Framework, Java Construction Framework), open source Distributed development framework Pebbl...

AI Technical Summary

Problems solved by technology

[0004] 1. At present, the commonly used distributed service systems in the industry are strongly dependent on a registration center. At the same time, for the high availability of the registration center, the registration center needs at least 3 nodes. Therefore, in order to maintain a highly available registration center, it increases the cost of operation and maintenance. The burden also increases a lot of hardware investment;

[0005] 2. The registration center is usually deployed in a cluster mode (many based on zookeeper implementation) to solve the problem of load balancing. Therefore, the address of the registration center needs to be maintained on the RPC server side. The strong dependency between the registration center and RPC services leads to distributed systems. The complexity of the

Images