When you are using more virtual memory than you have physical memory, the usual solution is to swap pages out to a hard disk.
What about compressing those pages, sticking them in a corner of physical memory, and decompressing them when they're needed again? This is what compcache does. It's a win for read times because zero latency + decompression time is still faster than hard disk latencies. And at a high level it makes sense because new computers typically have many processors but only one hard disk.