We are trying to integrate some sort of distributed cache into our system. We have two major requirements:
- High availability, i.e. automatic data replication to mirroring machines so that if one goes down, the data is around.
- Searchabilty/quaribility of the cache data, i.e. there is a need to do ranged searches.
- Optional: Returning the complete snapshot of the data stored.
- Optional: ability to persist cached data on periodic basis.
So far we have only SharedCache as a candidate because it allows at least searching using regex. That’s inherently slow though. SharedCache doesn’t support high availability. Windows App Fabric offers that but no search as far as I understand (if I am wrong that can be just awesome).
What are my options? Is that too much to ask? Should I give up and think about some sort of custom solution?
Definitely include Membase Memcached to the list of your considerations. It is very solid and stable. You can use the only “Memcached” part if you want but there is also NoSQL document database available (which of course can be used as a “cache persistence” too).