[jboss-jira] [JBoss JIRA] (JGRP-1680) RDMA based transport

Andrew Scully (Jira) issues at jboss.org
Sat Oct 27 06:12:00 EDT 2018


    [ https://issues.jboss.org/browse/JGRP-1680?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13653607#comment-13653607 ] 

Andrew Scully commented on JGRP-1680:
-------------------------------------

There is also a lot of discussion on the Open JDK nio-dev mailing list at the moment regarding first class support for rsockets within the JDK:
http://mail.openjdk.java.net/pipermail/nio-dev/2018-October/thread.html

Tracker:
https://bugs.openjdk.java.net/browse/JDK-8195160
https://bugs.openjdk.java.net/browse/JDK-8203434

Latest patch:
http://cr.openjdk.java.net/~ylu/8195160.15/

> RDMA based transport
> --------------------
>
>                 Key: JGRP-1680
>                 URL: https://issues.jboss.org/browse/JGRP-1680
>             Project: JGroups
>          Issue Type: Feature Request
>            Reporter: Bela Ban
>            Assignee: Bela Ban
>            Priority: Major
>             Fix For: 5.0
>
>
> Investigate whether an RDMA based transport makes sense.
> Advantages:
> * Speed, low latency (TCP/IP is bypassed entirely)
> * Low CPU usage
> Disadvantages:
> * JNI/C code
> ** Such a transport implementation would have to live outside of the JGroups repo
> ** Maintainability nightmare: the C code would also have to be ported to various OSes
> *** Investigate Java based libs (IBM's jVerbs) and C based libs (Apache Portable Runtime?)
> * High memory use, growing with cluster size: similarly to TCP, a 'group multicast' would involve N-1 sends. RDMA requires a Queue Pair (QP) for each destination. Each QP requires pinned memory (receive and send buffer), so each node would have to reserve (pin) N-1 memory buffers [1]
> ** OTOH, we may not use many group multicasts, e.g. with Infinispan's partial replication (DIST mode)
> * High cost of RDMA adapters, NICs and wiring: only a very small fraction of users would run such a transport.
> [1] http://www.hpcwire.com/hpcwire/2006-08-18/a_critique_of_rdma-1.html



--
This message was sent by Atlassian Jira
(v7.12.1#712002)


More information about the jboss-jira mailing list