hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Zephyr Guo (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-14490) [RpcServer] reuse request read buffer
Date Sun, 18 Oct 2015 16:24:05 GMT

    [ https://issues.apache.org/jira/browse/HBASE-14490?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14962485#comment-14962485
] 

Zephyr Guo commented on HBASE-14490:
------------------------------------

[~stack]
{quote}What is wrong w/ BoundedByteBufferPool ? {quote}
{{BoundedByteBufferPool.getBuffer()}} provide a buffer, but it can't assure that size is enough.For
example, if dataLen=10 and {{getBuffer()}} return a buffer that size less than 10, we have
to allocate again.Patch-11 maintain a dynamic buffer in {{Connection}}.I think BufferPool
is not different with Patch-11.I do some test and it seems Patch-10 is better.Patch-10 use
a fixed buffer for most small request(such as get operation).

YCSB test result.
{code}
*No reusable buffer:*
[OVERALL], RunTime(ms), 18583.0
[OVERALL], Throughput(ops/sec), 1614.3787332508207
[INSERT], Operations, 30000.0
[INSERT], AverageLatency(us), 563.6561

[OVERALL], RunTime(ms), 7997.0
[OVERALL], Throughput(ops/sec), 1250.4689258471926
[READ], Operations, 10000.0
[READ], AverageLatency(us), 629.6487

*Patch-11:*
[OVERALL], RunTime(ms), 18441.0
[OVERALL], Throughput(ops/sec), 1626.8098259313485
[INSERT], Operations, 30000.0
[INSERT], AverageLatency(us), 540.7597

[OVERALL], RunTime(ms), 7668.0
[OVERALL], Throughput(ops/sec), 1304.1210224308816
[READ], Operations, 10000.0
[READ], AverageLatency(us), 620.7383

*Patch-10(3K size):*
[OVERALL], RunTime(ms), 17313.0
[OVERALL], Throughput(ops/sec), 1732.8019407381737
[INSERT], Operations, 30000.0
[INSERT], AverageLatency(us), 526.9979333333333

[OVERALL], RunTime(ms), 7516.0
[OVERALL], Throughput(ops/sec), 1330.4949441192123
[READ], Operations, 10000.0
[READ], AverageLatency(us), 604.255
{code}

Seems like anoop said older version is better...
stack, what's your suggestion?


> [RpcServer] reuse request read buffer
> -------------------------------------
>
>                 Key: HBASE-14490
>                 URL: https://issues.apache.org/jira/browse/HBASE-14490
>             Project: HBase
>          Issue Type: Improvement
>          Components: IPC/RPC
>    Affects Versions: 2.0.0, 1.0.2
>            Reporter: Zephyr Guo
>            Assignee: Zephyr Guo
>              Labels: performance
>             Fix For: 2.0.0, 1.0.2
>
>         Attachments: ByteBufferPool.java, HBASE-14490-v1.patch, HBASE-14490-v10.patch,
HBASE-14490-v11.patch, HBASE-14490-v2.patch, HBASE-14490-v3.patch, HBASE-14490-v4.patch, HBASE-14490-v5.patch,
HBASE-14490-v6.patch, HBASE-14490-v7.patch, HBASE-14490-v8.patch, HBASE-14490-v9.patch
>
>
> Reuse buffer to read request.It's not necessary to every request free buffer.The idea
of optimization is to reduce the times that allocate ByteBuffer.
> *Modification*
> 1. {{saslReadAndProcess}} ,{{processOneRpc}}, {{processUnwrappedData}}, {{processConnectionHeader}}
accept a ByteBuffer instead of byte[].They can move {{ByteBuffer.position}} correctly when
we have read the data.
> 2. {{processUnwrappedData}} no longer use any extra memory.
> 3. Maintaining a reused ByteBuffer in each {{Connection}}.
> a) Size is dynamic to fit specific client.
> b) If size deviate far from average size, we replace a new one.
> c) Using a SoftReference to reference the buffer.



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

Mime
View raw message