I always thought MemSQL was a strong contender with the commitment to the MySQL wire protocol and SQL. The addition of a column store for larger than memory storage also seemed like a really good way to balance the concerns of speed of access vs storage cost.
I always wished we had taken a few pages from their playbook when I was @ VoltDB. Including ignoring BEGIN and END transaction :-)
This page: http://blog.memsql.com/powerstream-demo/ mentions:
"Real-time notifications push to the UI based on a `select *` query from an events table, which scales up to 2 million records.". But the docs don't really talk about doing things like that. Anymore info on these live queries?
In MemSQL Powerstream, we run queries against MemSQL for events which return in a couple of milliseconds. These queries are run continuously thus they are able to provide the front-end with events as they happen. At this time we do not have in-MemSQL live query support, but given enough interest it could be something we consider.
Props. May not be a tool I'm personaly likely to use (my colleagues and I work on a db meant for a very different work load in a perpendicular space), but I'm always in favor of more investment in real tech, even if it's c++ heavy :p
Also enjoyed meeting Eric in NYC a few years ago. Props!
I am working on a greenfield project right now utilizing MemSQL and so far its pretty solid. We haven't gone to production yet, so no concrete results to report.
The cool thing about it, compared to other in-memory tools, is that its communications protocol is identical to MySQL, therefore MySQL Workbench, MySQL Cli tools, all tools work with MemSQL out of the box.
I seriously looked at Redis and SQL Server 2016 in-memory DB. Redis was rejected because it isn't really a database. And SQL Server because it can't scale out, only up.
MemSQL was easy to install and do a quick test. Plus, you can just fire up a MemSQL appliance on Azure.
For MemSQL folks here...implement stored procedure support please. I would love to move all legacy apps I have over to MemSQL but I can't because of this.
I always wished we had taken a few pages from their playbook when I was @ VoltDB. Including ignoring BEGIN and END transaction :-)