Very fast GZIP compression by means of content addressable memories

Loss-free data compression is a common technique in various applications. Approaches that demand for real-time compression in the communication area aim a higher utilization of limited storage space or scarce bandwidth. While existing on-the-fly compression solutions in most of the cases consume a high portion of the system processor resources, this paper points out a new and unique compression scheme by means of dedicated hardware to disburden the main processor almost completely from this very extensive task. This scheme accelerates many compression standards that base on a longest match searching. GZIP is an algorithm that highly benefits from this acceleration and an example use is described in more details. This scheme is shown in conjunction with an application to accelerate Internet accesses, but can be applied to many other fields where very fast data compression is useful.