Ultra low latency in cluster application

[David Alves]

Hi Frederic
I thought that might be the problem too :). But this happens in different
vms and operating systems. My dev pc (where it works perfectly very high
throughput and low very low latency) is a mac on sun java 1.6(64bit). My
small test setup @home (where the problem also happens) is my mac + a ubuntu
running sun java 1.6 (32bit) too. In the cluster I have multiple configs,
some of the machines are pentiumD 3.0 (1) and some are core2duo (2) they all
run centOs. All the 1's are running sun java 32bit and at first all the 2's
were running iceTea jdk 7, 64 bit. Having 1 as server 1 as client, the
throughputs are not that great but in singe thread the latency is stable
(not in multithread). Running 1 as server 2 as client the latency becomes
unstable. I tryed changing the vm from iceTea to sun java but it still
happens.
As you can see I think the diverse environments rule out the specific jdk
reason, but I could be wrong. Any ideas?

Regards
David


On Wed, Feb 25, 2009 at 8:39 AM, Frederic Bregier <fredbregier at free.fr>wrote:

>
> Hi David,
>
> Just a quick question : what kind of servers do you use?
> (OS and level, JDK editor, level and 32 or 64bits)
> It may be not related, but I observed already some problem on a buggy JDK
> but not with the same behaviour as yours.
>
> Frederic
>
>
> -----
> Hardware/Software Architect
> --
> View this message in context:
> http://n2.nabble.com/Ultra-low-latency-in-cluster-application-tp2381448p2382588.html
> Sent from the Netty User Group mailing list archive at Nabble.com.
>
> _______________________________________________
> netty-users mailing list
> netty-users at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/netty-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/netty-users/attachments/20090225/1fdd5041/attachment.html