Fri, 15 May 2015 14:48:17 GMT
omething reminiscent of Google's infrastructure. At one point, they wrote about a 200ms turnaround goal for all search engine queries. They achieve it by keeping their entire index in RAM, sharded across machines.
> . . . the two primary reasons Redis sucks as a primary store . . .
IMO the reason is lack of an explicit schema and tool for migrations, something that makes all development with NoSQL difficult.