On 14 Oct 2010, at 08:32, Galder ZamarreƱo wrote:
What exactly do you guys mean by 'native'? Something like
what HornetQ guys did? Or your talking more about adapting an existing cache store impl
library to suit our type of ops?
I would actually like to see a LibAIO based impl for Linux. Similar in some ways who what
HornetQ has, but with some crucial differences: HQ uses this as a file journal, which is
sequentially accessed and always appended to. We would need random access, and most
likely some sort of hash/bucket system to ensure quick lookup on keys.
But I must stress that this cannot be high prio for now - we could spend a few spare
cycles and sketch up a design on the wiki, or maybe even invite designs from the
community, and see who contribs impls, for now.
Cheers
Manik
--
Manik Surtani
manik(a)jboss.org
Lead, Infinispan
Lead, JBoss Cache
http://www.infinispan.org
http://www.jbosscache.org