[jbosscache-dev] Re: [JBoss JIRA] Resolved: (JGRP-65) Performancetests and publish results

[Bela Ban]

We're *not* talking about using TCP as default, we're talking about
- looking at *all* configs in 
jgroups-multiplexer.sar/META-INF/multiplexer-stacks.xml
- optimizing those
- remove unneeded ones and
- add new ones (e.g. one without FC and message bundling for synchronous 
RPCs)

Brian Stansberry wrote:
> If what we're talking about here is switching to a TCP-based stack as
> the default, I'd be concerned about doing that until
> http://jira.jboss.com/jira/browse/JGRP-130 is resolved.
>
> If we're just talking using the optimized configs for each protocol, yes
> definitely.
>
> jbosscache-dev-bounces at lists.jboss.org wrote:
>   
>> Yep, +1
>>
>>
>>     
>>> In that case, let's complete the JIRA Brian created and then simply
>>> copy over the config to JBossCache
>>>
>>> Manik Surtani wrote:
>>>       
>>>> I was also talking about the default XML files shipped with JBC for
>>>> standalone use. 
>>>>
>>>> Multiplexer related changes would only be for usage within AS. --
>>>> Manik Surtani
>>>>
>>>> Lead, JBoss Cache
>>>> JBoss, a division of Red Hat
>>>>
>>>> Email: manik at jboss.org
>>>> Telephone: +44 7786 702 706
>>>> MSN: manik at surtani.org
>>>> Yahoo/AIM/Skype: maniksurtani
>>>>
>>>>
>>>> On 25 Aug 2006, at 13:18, Bela Ban wrote:
>>>>
>>>>         
>>>>> Yes - Brian has a JIRA task open regarding this, for the
>>>>> Multiplexer. Meaning, you wouldn't be affected
>>>>>
>>>>> Manik Surtani wrote:
>>>>>           
>>>>>> Would this mean a prudent change in the default configs shipped
>>>>>> with JBC/JBossAS? --
>>>>>> Manik Surtani
>>>>>>
>>>>>> Lead, JBoss Cache
>>>>>> JBoss, a division of Red Hat
>>>>>>
>>>>>> Email: manik at jboss.org <mailto:manik at jboss.org>
>>>>>> Telephone: +44 7786 702 706
>>>>>> MSN: manik at surtani.org <mailto:manik at surtani.org>
>>>>>> Yahoo/AIM/Skype: maniksurtani
>>>>>>
>>>>>>
>>>>>> On 25 Aug 2006, at 12:39, Bela Ban wrote:
>>>>>>
>>>>>>             
>>>>>>> Ben Wang wrote:
>>>>>>>               
>>>>>>>> Bela,
>>>>>>>>
>>>>>>>> Very good documentation and good numbers. I think once
>>>>>>>> threadless stack is ready in 2.5, we will see dramatic
>>>>>>>> performance boost on JBoss Cache level as well!
>>>>>>>>                 
>>>>>>> Absolutely ! Parellel processing will definitely help both
>>>>>>> JBossCache and clustering (ClusterPartition).
>>>>>>>
>>>>>>>               
>>>>>>>> So looks like TCP performs better than UDP now. Is this the
>>>>>>>> first time that we have seen this?
>>>>>>>>                 
>>>>>>> No, that's always been the case for those smaller clusters. It
>>>>>>> will be interesting to see what happens on a 64 node cluster
>>>>>>> though. 
>>>>>>>
>>>>>>> --Bela Ban
>>>>>>> Lead JGroups / Manager JBoss Clustering Group JBoss - a division
>>>>>>> of Red Hat
>>>>>>>               
>>>>> --Bela Ban
>>>>> Lead JGroups / Manager JBoss Clustering Group JBoss - a division
>>>>> of Red Hat 
>>>>>
>>>>>           
>>>>         
>>> --
>>> Bela Ban
>>> Lead JGroups / Manager JBoss Clustering Group JBoss - a division of
>>> Red Hat 
>>>
>>>       
>> _______________________________________________
>> jbosscache-dev mailing list
>> jbosscache-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jbosscache-dev
>>     
>
>
>
> Brian Stansberry
> Lead, AS Clustering
> JBoss, a division of Red Hat
> Ph: 510-396-3864
> skype: bstansberry
>
>   

-- 
Bela Ban
Lead JGroups / Manager JBoss Clustering Group
JBoss - a division of Red Hat