Sorry I think I might have lead you in the wrong direction there. I'm talking about deep in the bowels of the terracotta server where we process transactions the place where we actually apply deltas to the state of a some particular clustered object (one of a CDM sub maps in this case) -- that application can only be done by single thread. The data being applied here might very well be folded and batched mutations made by many threads in various VMs using differing locks mind you. This detail is probably not nearly as important as the first reason which has to do with distribution in a array of terracotta servers.
-----Original Message----- From: Sergio Bossa [mailto:sergio.bo...@gmail.com] Sent: Thursday, April 07, 2011 3:12 PM To: tc-dev@lists.terracotta.org Cc: Tim Eck Subject: Re: [tc-dev] Purpose of concurrency level in ConcurrentDistributed(Server)Map On Thu, Apr 7, 2011 at 11:29 PM, Tim Eck <t...@terracottatech.com> wrote: > The other reason I think has to do with getting a more mutation > concurrency. At the lowest level in the terracotta server a clustered > object can only have changes applied to it by a single thread. Thanks for your answer, Tim. So, at the lowest level, a CDM can be concurrently mutated by only a single thread, regardless of the mutating key and the higher level lock acquired on that key? Am I getting it right? Isn't it like a global lock? -- Sergio Bossa http://www.linkedin.com/in/sergiob _______________________________________________ tc-dev mailing list tc-dev@lists.terracotta.org http://lists.terracotta.org/mailman/listinfo/tc-dev
