Because Redis requires the entire dataset to reside in memory.
It doesn't: http://code.google.com/p/redis/wiki/VirtualMemorySpecificati... .. Well, in the classical "in memory" sense. Virtual memory is a sort of memory.. :-)
It doesn't: http://code.google.com/p/redis/wiki/VirtualMemorySpecificati... .. Well, in the classical "in memory" sense. Virtual memory is a sort of memory.. :-)
Similar problems with MongoDB. Anytime your indexed data starts hitting disk, kiss your site goodbye.
Sharding seems like the proper workaround though, so maybe you could take the same approach of flockdb and shard the redis social graph.
the performance degrades to a point where it'll be too
slow to do anything
You can get boxes with a ton of RAM these, days, though, and if you go beyond it, you can always shard/distribute the database, like you mention.