One question I have is the cold start problem. How can I ensure dropping in a new peer is not going to have a large negative effect on response times? With memcache, you can just prewarm a new node or have clients only round-robin it a few times per request to warm it up. It seems like pre-warming here is going to be more cumbersome since it's not a simple k-v store but will require you to pre-emptively run queries to get there. (Similar to Lucene.)

Edit: Rich's response here:

http://blog.fogus.me/2012/03/05/datomic/comment-page-1/#comm...

Seems to imply that non-cached performance won't be so bad anyway. Looking forward to seeing some benchmarks.