Cache updating method suitable for divergent data

Abstract

The invention discloses a cache updating method suitable for divergent data. According to the method, double cache queues are designed for the problems that the cache hit rate of divergent data is lowand the divergent data is prone to being affected by historical data, one cache queue A conducts cache updating mainly according to the data access condition, the other cache queue B conducts cache updating according to the data updating condition, and the capacity of the cache queue A is larger than that of the cache queue B. The cache queue A maintains an access mark, and when the space is full, capacity expansion is carried out or some marked cache items are deleted. And the cache queue B maintains the updating time of the cache items, and deletes the cache item which is updated first whenthe space is full. According to the method, the problem that a traditional cache updating algorithm is low in cache hit rate under the condition of data divergence is solved, meanwhile, the conditionof historical data pollution is avoided, the method is sensitive to data access changes and can rapidly adapt to the data access changes, and the high cache hit rate is achieved through a small memory space.

Description

technical field [0001] The invention relates to a cache update method, in particular to a cache update method in the case of divergent data. Background technique [0002] With the development of the Internet, the number of Internet users has increased dramatically. The concurrency of many systems has also increased, and the operation of the database has become the concurrency bottleneck of the system. The research found that most of the data requests are query requests, so the cache can be used to avoid occupying database resources, and the database data is requested only when the cache misses. In actual production, not only the server will use cache, but many clients also have corresponding cache mechanisms. The application effect of caching on the server side is better, which greatly improves the concurrency of the system. However, the cache capacity cannot be expanded without limit. When a large amount of data is added to the cache, a trade-off needs to be made accordi...

AI Technical Summary

Problems solved by technology

In practice, LFU is very common. It is similar to LRU and is also used to process hot data, but usually has a higher hit rate.

However, LFU also has certain problems. It cannot make adjustments in time according to the data access situation. Compared with other algorithms, it takes longer to adapt to the new data access mode, because the "historical data" in the cache needs a longer period of time. cycle to change

[0007] However, when the data access is relatively divergent and there is no hot data effect, the hit rate of the traditional cache update method will drop significantly

At the same time, when the data access mode changes, many traditional cache update methods cannot quickly adapt to the new data access mode

Images