Data compression technique for PC communications

Abstract

An adaptive technique is described for compressing data that are flowing through a communications channel, being transferred to and from a personal computer using a modem to connect to a remote computer or database service. The objective is to reduce the transmitted data to close to five bits for each character of input data in typical ASCII text files with an algorithm that is simple enough that an eight-bit microprocessor, such as a Z-80, can handle a 19200-b/s asynchronous channel and still be able to handle, in addition, X.25/LAPB error-control protocol functions. This allows the data to be transmitted from modem to modem at 9600 b/s with synchronous transmission using the fast-turnaround ping-pong protocol developed for the Hayes V-series 9600 modems, with the user connected to the modem at 19200 b/s. The technique described meets those objectives for most text files and for most Lotus worksheet files as well. Less effective compression is realized when more random data are transmitted, such as executable binary files; however, the compression ratio is always close to the theoretical limit given by information theory for character-by-character compression. As an example of the effectiveness of this technique, it reduced the ASCII text file of a draft of this paper from 10928 to 6583 bytes, or to 60% of the original size. This corresponds to 4.8 bits per character.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>