Parallel Algorithms for On-line Dynamic Data Compression

We consider parallel algorithms for on-line (realtime) data compression that employ dynamic (adaptive) “learning” by textual substitution (e.g., techniques studied in the past by this author, Lempel and Ziv, Welch, Miller and Wegman, etc.). Our model of computation is a systolic pipe, where processors are arranged in a linear array (each processor is connected only to its left and right neighbors). We first consider an array of processors that is large enough to accommodate one dictionary entry per processor. We then discuss the using an existing parallel machine where the number of available processors may be limited.