Sub-micron technology development and system-on-chip (Soc) design - data compression core

Abstract

Data compression removes redundancy from the source data and thereby increases storage capacity of a storage medium or efficiency of data transmission in a communication link. Although several data compression techniques have been implemented in hardware, they are not flexible enough to be embedded in more complex applications. Data compression software meanwhile cannot support the demand of high-speed computing applications. Due to these deficiencies, in this project we develop a parameterized lossless universal data compression IP core for high-speed applications. The design of the core is based on the combination of Lempel-Ziv-Storer-Szymanski (LZSS) compression algorithm and Huffman coding. The resulting IP core offers a data-independent throughput that can process a symbol in every clock cycle. The design is described in parameterized VHDL code to enable a user to make a suitable compromise between resource constraints, operation speed and compression saving, so that it can be adapted for any target application. In implementation on Altera FLEX10KE FPGA device, the design offers a performance of 800 Mbps with an operating frequency of 50 MHz. This IP core is suitable for high-speed computing applications or for storage systems.