带智能预测缓存的物联网标识解析器

针对现有物联网标识解析存在的性能缺陷, 提出一种带智能预测缓存的物联网标识解析器, 以DNS为基础, 支持各种物联网标识编码类型的解析, 并具有智能预测的缓存功能。预测时采用关联规则方法对日志数据进行挖掘, 根据当前的查询请求预测下一个可能解析的标识, 提前解析, 缓存解析结果。同时结合TTL策略以及LRU策略作为缓存置换方法, 具有高效性和时效性。实验表明, 该解析器具有较高的命中率和较小的响应时间, 且当缓存大小为62时获得最优性能。

IoT identifier resolver with intelligent predictive cache

In order to improve the performance of the traditional IoT identifier resolving methods, this paper proposed an IoT identifier resolver with intelligent predictive cache. The resolver was based on DNS, and supported multi-coding resolving and had a cache with the function of intelligent prediction. Mining the log data of the resolver with association rules, the resolver could predict the identifier to resolve based on the current request and pre-resolve and cache the coming requested code. The replace strategy of the resolver combined the TTL strategy with the LRU strategy at the same time, which provided with high performance and timeliness. The experimental results show that the resolver has a higher hit rate and quicker response and get the best performance when setting the cache size as 62.