FriendFeed actually used a similar mechanism with MySQL for some schemaless data that they knew they wouldn't need to index. "Entity bodies are stored as zlib-

FriendFeed actually used a similar mechanism with MySQL for some schemaless data that they knew they wouldn't need to index. "Entity bodies are stored as zlib-compressed, pickled Python dictionaries." http://backchannel.org/blog/friendfeed-schemaless-mysql http://backchannel.org/blog/friendfeed-schemaless-mysql