[infinispan-issues] [JBoss JIRA] Commented: (ISPN-359) Provide mechanism for users to control data locality when using DIST

[Vladimir Ralev (JIRA)]

    [ https://jira.jboss.org/jira/browse/ISPN-359?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12524815#action_12524815 ] 

Vladimir Ralev commented on ISPN-359:
-------------------------------------

By Buddy Replication I meant the DIST mode. I thought it is still called that way. Anyway, I will see what Brian says about mod_cluster. Other similar products use configurable affinity key independently in the cache and in the load balancers. In Mobicents we use pluggable algorithms for example and we will probably try to teach it to cooperate with this feature deterministically to cover some telco cases (some telco protocols dont have cookies  or redirects to hunt down the right node).

But what is still unclear on this topic is the rehashing/rebalancing. If a node fails is everything completely re-hashed to a new location (along with the huge network overhead of moving the chunks around over the network) or is there some effort to minimize the state transfer?

> Provide mechanism for users to control data locality when using DIST
> --------------------------------------------------------------------
>
>                 Key: ISPN-359
>                 URL: https://jira.jboss.org/jira/browse/ISPN-359
>             Project: Infinispan
>          Issue Type: Feature Request
>          Components: Distributed Cache
>    Affects Versions: 4.0.0.Final
>            Reporter: Tero Heinonen
>            Assignee: Manik Surtani
>             Fix For: 4.1.0.BETA1
>
>
> Programmable affinity either through:
>  
> 1) statically (e.g. annotating a certain field as the key for consistent hashing)
> - in our case this cannot be based on the hashCode() as the affinity requirement and object identity are not directly related
>  
> 2) dynamically (e.g. providing a overridable/annotatable consistentHashCode() function)
>  
> This would help to assign entities (in multiple Caches), which are needed together in the same nodes (to avoid remote lookups)

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira