Distributed Caching System Based On Semilattice Data Structure

As the popularity of smart phones and the increasing number of Internet platforms,the server back-end often faces a large number of high concurrent accesses nowadays.The processing speed of the server front-end is out of proportion to the growth of user visits.Considering the user's experience and the relief of server-side pressure,it is necessary to establish a distributed cache system with high availability.It can not only enable users to get timely feedback from the system without overtime when using software such as APP,but also enable common data to be stored in the cache system to reduce the pressure on the server.Therefore,a distributed cache system that maintains high availability is very suitable for today's network access scenarios.Starting from the pressure of the current hot Internet business on the server side,we need to design a distributed cache system with high availability to improve the user experience and relieve the pressure of the server.The basic functions of the system are only access and query.Among them,the information transfer mechanism uses Zeromq message queue with fast speed and various modes.The client agent part matches the destination for each transmitted data;The front-end of the server complete the collection of data from various places and the distribution of requests.The Actor model part mainly performs specific access functions.Replica nodes maintain the final consistency of data and memory management.The design of data format mainly adopts the structure of semi-lattice attribute.The semi-lattice structure has three characteristics:relevance,interchangeability and idempotency.It is reflected in the distributed cache system that nodes are not sensitive to problems such as duplication,disorder and grouping when receiving data.Thus,it can be directly stored in the copy quickly.This data format can achieve final consistency.