Lossless compression of industrial time series with direct access

Abstract

The new opportunities generated by the data-driven economy in the manufacturing industry have caused many companies opt for it. However, the size of time series data that need to be captured creates the problem of having to assume high storage costs. Moreover, these costs, which are constantly growing, begin to have an impact on the profitability of companies. Thus, in this scenario, the need arises to develop techniques that allow obtaining reduced representations of the time series. In this paper, we present a lossless compression method for industrial time series that allows an efficient access. That is, our aim goes beyond pure compression, where the usual way to access the data requires a complete decompression of the dataset before processing it. Instead, our method allows decompressing portions of the dataset, and moreover, it allows direct querying the compressed data. Thus, the proposed method combines the efficient access, typical of lossy methods, with the lossless compression.