This is of course an interesting option - well, we could also leverage indexed disk cache as L3 - and steal indexing and recycling techniques. Merging efforts with JCS could bring to the most complete java cache solution in the OS world. I don't have any idea about performance of JCS and its lateral cache, but it's probably a bit outdated and could benefit from a technology refresh (NIO, fast serialization - see Apache Avro as an example)
On Mon, Nov 7, 2011 at 10:30 PM, Mir Tanvir Hossain < [email protected]> wrote: > JCS already has distributed caching option. So, instead of re-implementing > the distributed cache, we can integrate tightly with JCS, and become its > off-heap plugin. So, JCS could be the L1 cache, and DirectMemory could be > the L2 cache. Just an idea. > > -Mir > > On Mon, Nov 7, 2011 at 12:30 PM, Ioannis Canellos <[email protected]> > wrote: > > > I was looking at the jira issues and DIRECTMEMORY-13 - Make some > > investigation about going > > distributed<https://issues.apache.org/jira/browse/DIRECTMEMORY-13> > > caught > > my attention. > > > > I've read about some of the initial thoughts about reusing features of > > existing Caches (Hazelcast or Terracotta) but I don't like this idea so > > much. > > We could reuse some of the ides that existing projects use, but I think > > that we should either build it on our own, or use a solution not coupled > to > > an existing cache. > > > > So what are your thought about making direct memory distributed? > > Any thoughts about the consistency model, discovery etc? > > > > -- > > *Ioannis Canellos* > > * > > FuseSource <http://fusesource.com> > > > > ** > > Blog: http://iocanel.blogspot.com > > ** > > Apache Karaf <http://karaf.apache.org/> Committer & PMC > > Apache Camel <http://camel.apache.org/> Committer > > Apache ServiceMix <http://servicemix.apache.org/> Committer > > Apache Gora <http://incubator.apache.org/gora/> Committer > > Apache DirectMemory <http://incubator.apache.org/directmemory/> > Committer > > * > > >
