Research Of The Fast Lossless Compression Algorithms On Embedded Platform

With the rapid development of the Internet and mobile communication, the amount of data storage and transmission of smart devices shows an explosive growth trend. The processing procedure of massive data has brought about more storage and transmission problems under the existing conditions of the hardware resources and network bandwidth. To make the embedded devices in a certain scenario both meet real-time requirement, but also use the limited hardware resources and network bandwidth efficiently, various types of data should be compressed.Based on lossless compression theory, in this paper, the current lossless compression technique is introduced. We focus on research work of fast lossless compression algorithm in certain scenarios. By analysis of impact factors on performance of DEFLATE algorithm, this paper presents parameters configuration of DEFLATE algorithm in embedded system and verifies the rationality. And also we propose fast lossless compression algorithm in file system. The main content of this paper is summarized as follows.1) In the first part, we analyze DEFLATE algorithm, which includes a detailed description about the Huffman entropy encoding and algorithm of searching the match data. Then we use the Galgary Corpus data set to test and analyze the influence on the compression and time efficiency of the DEFLATE algorithm from several aspects, including match algorithm, the length of hash chain, match length and Huffman coding. By performance analysis of DEFLATE algorithm, this paper presents parameters configuration of DEFLATE algorithm in embedded system and verifies the rationality. Experimental results show that compared with using default parameters, compression speed of DEFLATE algorithm using tuning parameters given in this paper improves by 33.52%, while the compression ratio only redeces by 0.89 %.2)In this paper, the fast lossless compression algorithm is proposed for data unit in file system, and this allows for quick access to the data. Based on LZO algorithm, structure characters is considered by correlation analysis on data in file system. Aiming at improving encoding speed in comparison to LZO algorithm, we design a new method to search the match data and encoding format for data unit in file system. Experimental results show that the compression/decompression speed of algorithm proposed in this paper increases by more than 25% and 32% repectively, while our algorithm drops by 0.81% on compression ratio performance in comparison to LZO algorithm.Through this paper research, it provides evidence of parameters configuration of DEFLATE algorithm in certain scenarios. And the fast lossless compression algorithm based on file system is proposed, it can be used in compressible file system of embedded device to meet real-time requirement, also saving storage space and providing better access speed for data in embedded file system.