Server Cache Management Mechanism Based on Bayesian Probability and Comprehensive Value

Aiming at the problems of low hit ratio and increasing response delay caused by exponential growth of Internet users, a cache management mechanism suitable for server local cache is proposed. The Bayesian network is used to calculate the probability of data being accessed continuously, and the decision whether to cache data is made based on the probability. After the data is cached, the number of data being accessed within a unit time is counted to obtain the data activity. When the data in the cache area needs to be recovered, the Newton cooling law is used to calculate the proximity value of cached data to protect the new data. The CRITIC method is adopted to calculate the continuous access probability of cache data, the activity and the weight of adjacent value index, and the comprehensive value is obtained by linear weighting. When the cache area capacity is insufficient, the cache data with the comprehensive value lower will be removed. The experimental results show that compared with the traditional cache management mechanisms LFU and FIFO, the proposed cache management mechanism improves the cache hit ratio by 20.96% and 32.46%, respectively, and reduces the server response delay by 24.94% and 31.74%.