Zedis distributed type buffer method

Abstract

The invention provides a zedis distributed type buffer method. The zedis distributed type buffer method comprises the following steps of enabling a server judging module to read the complete information required by a redis server cluster of a server cluster monitoring module; enabling the server judging module to send the read information to a client, and enabling the client to read the complete information required by the redis server cluster; extracting unique information from the complete information of each physical node corresponding to other physical nodes, wherein the information comprises an ip (Internet protocol) address and a port number; enabling the client to generate a plurality of fixed keys according to the received information, generating a plurality of hash codes by a load balance core type ConsistentHash algorithm, mapping all hash codes generated by the same information into the same physical node, and filling a ConsistentHash core variable mapping table to form a hash ring by mapping the hash codes into the physical node. The zedis distributed type buffer method has the advantages that by utilizing the minimum configuration, the high availability and automatic control of the cluster are realized, and the data transfer, load balance, and dynamic and stable properties are realized.

Description

technical field [0001] The invention relates to the technical field of data caching, in particular to a distributed caching method. Background technique [0002] The current data caching schemes include redis cluster scheme and twitter scheme. [0003] The redis cluster solution, due to its relatively new launch time, was not yet mature and used on a large scale when our solution was proposed. It provides a hash algorithm for pre-allocated buckets, which is not as flexible as the consistent hash algorithm. Failover requires Assign manually. [0004] The twitter solution, which acts as a proxy directly at the redis protocol layer, can be flexibly configured and supports automatic algorithms such as consistent hashing, but it is a static distributed cluster that does not support automatic failover of redis nodes, and this The failover of the broker itself also needs to be done by developers themselves. [0005] Combining redis cluster and twitter to start a business, in add...

AI Technical Summary

Problems solved by technology

Solution 4 is relatively heavy, and it is more difficult to integrate our internal platform

Images