I want to compress a large nuber of files that have near identical headers, and also some data, however I do not wish to archive them, nor do I wish to zip them individually(because the copression ratio would be much higher if substitutions of similar blocks could be done using a single table). Does a compression method exist to do this already, or should I implement it myself.

Note: Don't say "Disk space is cheap", because I may want to use this on an embedded system.