Tugela cache

I came across Tugela Cache by way of the High Scalability weblog.

It is a disk-based implementation of memcached, a memory based caching system.

I actually implemented something very similar to that while at Feedster to serve as a central cache for the search engine. I used the memcached protocol, but my disk-based implementation was a little different since I used SHA1 keys to identify cache files, hence I could use a multi-level directory based approach to storing the files.

In the event I never used it because it created a single point of failure as well as a choke point for the search engine, which is contrary to good scaling practice.