An Intelligent Time Series Storing and Inference Engine Implementation with Focus on Performance and High Levels of Abstraction

This paper covers several aspects of intelligent time series database implementation. It also includes the description and analysis of a symbolic time series representation scheme. The paper focuses on various indexing and parallelization approaches in conjunction with actual backend storage engines. Special emphasis is made on identifying the problem of combining simple queries with time series pattern search and retrieval requests and finding a solution to this problem. The paper also considers query definition and provides the general architecture of a time series database with data mining capabilities.