hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sandy Pratt (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-8691) High-Throughput Streaming Scan API
Date Thu, 13 Jun 2013 20:49:20 GMT

    [ https://issues.apache.org/jira/browse/HBASE-8691?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13682705#comment-13682705
] 

Sandy Pratt commented on HBASE-8691:
------------------------------------

I noticed when doing some more testing that I'm returning only a partial Result in the stream
case, but the full Result in the control cases.  That means the stream result doesn't have
to transfer the rowkey (x3) and a couple of timestamps.  When I correct the test to return
the full Result, using DataOutputBuffer/DataInputBuffer to serialize and deserialize, the
stream case is about 2x faster than the RPC cases (rather than 4x faster).  I think it's still
worth looking at for a 2x speedup.

Unfortunately, I can't figure out how to make Hive work with an event driven interface.  Since
my customers use Hive primarily, I might have to put this on the back burner for now.  I hope
somebody finds it useful.
                
> High-Throughput Streaming Scan API
> ----------------------------------
>
>                 Key: HBASE-8691
>                 URL: https://issues.apache.org/jira/browse/HBASE-8691
>             Project: HBase
>          Issue Type: Improvement
>          Components: Scanners
>    Affects Versions: 0.95.0
>            Reporter: Sandy Pratt
>              Labels: perfomance, scan
>         Attachments: HRegionServlet.java, README.txt, RecordReceiver.java, ScannerTest.java,
StreamHRegionServer.java, StreamReceiverDirect.java, StreamServletDirect.java
>
>
> I've done some working testing various ways to refactor and optimize Scans in HBase,
and have found that performance can be dramatically increased by the addition of a streaming
scan API.  The attached code constitutes a proof of concept that shows performance increases
of almost 4x in some workloads.
> I'd appreciate testing, replication, and comments.  If the approach seems viable, I think
such an API should be built into some future version of HBase.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

Mime
View raw message