4 ms·
If "key value" is important, have you considered Redis? If things like Cassandra and HBase are the Lisp or Haskell of the database world, Redis is the assembler
by petercooper 16y ago
If "key value" is important, have you considered Redis? If things like Cassandra and HBase are the Lisp or Haskell of the database world, Redis is the assembler - it's crazy fast but you need to take care of your own underlying data structures (beyond strings, lists, and two types of set).
- bravura 16y agoI guess the main question when considering Redis is: Does the entire database fit into memory?
- fictorial 16y agoRedis supports "virtual memory".
- riffraff 16y agothat would be "does the entire keyspace fit in memory?"
- ghotli 16y agoMy initial thoughts were for a Redis style database. It certainly won't live in memory. That's why I've been pushed towards a dynamo / bigtable approach.
- petercooper 16y agoThough, as has been raised a few times, that is not a restriction in Redis anymore.
- strlen 16y ago[Disclaimer: I am a committer on Project Voldemort] Redis is very cool (I really want to play with it - congrats to VMWare on hiring its author), but it is neither a disk-oriented system [EDIT: fixed per comment, was disk-persistent] nor (most importantly, IMO) a fully distributed system. See a comment in this blog post for great visual guide to these systems from these two axes (disk-orientation and distribution): http://pl.atyp.us/wordpress/?p=2929 http://pl.atyp.us/wordpress/?p=2929
- uggedal 16y agoRedis supports full disk persistence by adding the following values to your redis.conf: appendonly yes appendfsync always Yes, fsyncing after every operation will be slower than fsyncing every second or so.
- strlen 16y agoThanks, fixed it.
- petercooper 16y agoSure, but the OP specified: I need access to that data as quickly as possible. Writes are only going to occur quarterly. The emphasis is on speed and writes are rare, so being disk-oriented wouldn't provide any advantages.