That said it can be more than a little painful to watch it go through the normal growing pains associated with a promising project. […] When we learned that Redis VM wasn’t performing at scale in production environments it really spooked us. When antirez said he was looking down the barrel of implementing his own BTree (not even the best solution for modern storage backends) from scratch I started to get upset. Like angry upset. When the news started to float about the filesystem datastore (one file per key) I started to look for other solutions.
I’ve been hearing other people expressing concerns about both Redis disk persistency and Redis going distributed. But wouldn’t be a pity to have Redis’ data structures and accompanying operations on a single memory-only node?
Original title and link: Redis: Why Wait for Good Enough? (NoSQL databases © myNoSQL)