Time Series Data Compression

• Main Authors: BAMOUNI DOMINIQUE, 巴穆得
• Other Authors: Yuan, Shyan-Ming
• Format: Others
• Language: en_US
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/7t99x6

碩士 === 國立交通大學 === 電機資訊國際學程 === 106 === In recent years, numerous smart meters have been widely installed to aggregate time series engineering parameters over fields; it has led to problems of handling big data. The huge volumes of data need to be transmitted, stored, processed as well as retrieved. Storing and accessing these big data have become expensive in time, space and bandwidth. The aim of the study is to find a solution for the problems. One solution developed in the study is to compress/decompress the engineering parameters. The data format of the variables has three (03) portions: 128-bit Global Unique Identifier (GUID), 64-bit time stamp parameter, and 64-bit floating point value parameter. Three encoding/decoding algorithms have been applied and implemented. The approaches have reduced the original historical data size 40% off as well as the storage cost. The algorithms’ performances: the compression ratio, the saving percentage and the compression/decompression time and speed have been measured. The decompression process has been proved faster than the compression process based on the historical data.