I have currently have a web application that caches a large amount of data (several hundred thousand entries) in memory for quick lookup and then in SQL Server as a persistent cache. Basically the information consists of geocodes of addresses where the geocode is retrieved via a remote web service which takes time if needed to be called continuously rather than cached.
Would using RavenDB (or other suggestions) be a better way of caching this information in a persistent store in terms of both speed and memory? Currently the app is hosted on a server which SQL Server sucks most of the memory dry so I would like a more light weight solution.
RavenDB as a document database seems overkill to me if only used as a cache. What about redis as a value store? It could better match what you are after
RavenDb should be faster for reading than SQL. Not that writing is slow, but it's optimized for read operations as far as I know. So you should try it, it should be better performance-wise. I'm not sure about memory though.
If you want more info, you can always try the RavenDb Google Group. Ayende follows it, so he might answer himself.
来源：https://stackoverflow.com/questions/3350313/using-ravendb-as-a-persistent-cache