On 19 Mar 2011, at 17:02, Israel Lacerra wrote:

Guys,

Sorry. It's probably not a dev question... 

I'm writing a paper for my graduation (about Infinispan) as I said before. I'm having some doubts about what Infinispan really is. I watched some Manik presentations on internet and read a little articles. I saw Manik saying something like Infinispan be a in memory data base, and not just a distributed cache. And I read (http://java.dzone.com/articles/data-service-data-fabric) that Infinispan is something like Voldermort or Dynamo. And in some other sites, like http://nosql-database.org/, Infinispan is classified like a "Grid Database solution" (like Oracle Coherence) and somehow different of Voldermort. Reading Infinispan documentation and these and other articles, I not really sure what differs Infinispan and Voldermort.

I'm not really sure if I made myself clear, but I'm having some problems defining what is Infinispan (or what Infinispan wants to be in the end) (comparing with these other products).

Speaking of Voldemort specifically, they're both very similar in that they are both based on concepts from Dynamo.  A consistent hash based, distributed in-memory key/value store.  Application can differ though; since they're both in-memory, they make good distributed caches (like Oracle Coherence).  Naming and terminology gets a bit hazy from this point on, since with persistence, JTA, etc., it starts to make sense as a NoSQL database.

Hope this helps
Manik
--
Manik Surtani
manik@jboss.org
twitter.com/maniksurtani

Lead, Infinispan
http://www.infinispan.org