[ 
https://issues.apache.org/jira/browse/HBASE-13819?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14909002#comment-14909002
 ] 

stack commented on HBASE-13819:
-------------------------------

bq. So in client side we will continue with on heap only. 

We don't want to do offheap on client-side for same benefit?

bq. In server side let us make only off heap (for trunk)

Yes.

bq. Because the max direct buffer has to consider this capacity as well. We 
will keep these buffers in our pool. WIth max size of 2MB per buffer and twice 
the #handlers. (correct?) May be that will be a surprise for changes in 1.x 
versions?

Enable for 1.3 with a release note? It is a nice perf benefit. So, I suppose 
make it configurable, yeah. Thanks [~anoop.hbase]


> Make RPC layer CellBlock buffer a DirectByteBuffer
> --------------------------------------------------
>
>                 Key: HBASE-13819
>                 URL: https://issues.apache.org/jira/browse/HBASE-13819
>             Project: HBase
>          Issue Type: Sub-task
>          Components: Scanners
>            Reporter: Anoop Sam John
>            Assignee: Anoop Sam John
>             Fix For: 2.0.0
>
>         Attachments: HBASE-13819.patch
>
>
> In RPC layer, when we make a cellBlock to put as RPC payload, we will make an 
> on heap byte buffer (via BoundedByteBufferPool). The pool will keep upto 
> certain number of buffers. This jira aims at testing possibility for making 
> this buffers off heap ones. (DBB)  The advantages
> 1. Unsafe based writes to off heap is faster than that to on heap. Now we are 
> not using unsafe based writes at all. Even if we add, DBB will be better
> 2. When Cells are backed by off heap (HBASE-11425) off heap to off heap 
> writes will be better
> 3. When checked the code in SocketChannel impl, if we pass a HeapByteBuffer 
> to the socket channel, it will create a temp DBB and copy data to there and 
> only DBBs will be moved to Sockets. If we make DBB 1st hand itself, we can  
> avoid this one more level of copying.
> Will do different perf testing with changed and report back.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to