Form-Based Semantic Caching on Time Series

Time Series Databases Management System (TSMS) has been overcoming the Database Management Systems (DBMS) in storing vast amounts of data [33]. Nevertheless, TSMS only supports simple aggregate functions to analyze Time Series Data (TSD). Besides, to accelerate and save data transferring between clients and servers in the DBMS, semantic caching can be used. However, the semantic caching approach is not efficient because of not fully supporting aggregate functions in TSMS. Furthermore, the query result of TSD in the semantic caching technique could be huge for the in-memory database where the semantic caching technique is running on. A model-based compression can be used to compress data, reducing the data space in the in-memory database. In this paper, we present Form-based semantic caching for TSD system. The approach reduces both query result storing based on semantic caching technique and the data transfer between clients and servers. In particular, the approach accelerates up to 122 and 1.82 times the execution speed, comparing to the without cache and basic semantic caching approaches, respectively. On the public Reference Energy Disaggregation Data Set, the compression model ratio in the approach can be reached to 526.8:1.