In memory distributed caching vs In memory data grid

Question

What is the difference between "In memory distributed caching" vs "In memory data grid" ?

When do we use one over the other ie, what are the practical use cases for "In memory data grid" ?

Can you name few popular "In memory data grid" frameworks which are compatible with Java applications ?

Answer 1

In-memory distributed caching is just that - a cache that is distributed across different nodes. It makes data highly available to the application(s) using it. They're usually key/value stores and support the standard put/get operations along with abilities to partition, replicate or backup data.

An in-memory data grid is a distributed cache with a bit of computing power thrown in. Over the abilities of a distributed cache, it allows you to do distributed SQLs, co-located processing etc... plamb has given a good list of in-memory data grids.

Answer 2

With a In memory data grid you can do a in memory distributed cache. For example Oracle use Oracle Coherence to implement that kind of cache with Weblogic. So with this example I answered the last part of your question.

But it's the most expensive solution to just doing cache (money, memory, network, cpu) : In memory data grid is reliable more that you need to do cache, it can handle the real data so you don't need another backend.

If you need to do cache in memory distributed, solutions like EHCache, Memcached, Infinispan can do it. In fact almost Java EE application server give a solution to do it.