[
https://issues.apache.org/jira/browse/HBASE-9631?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13781276#comment-13781276
]
Andrew Purtell commented on HBASE-9631:
---------------------------------------
I looked over the referenced Cassandra JIRA. A test of theirs showed a small
increase in the false positive rate for their bloomfilters:
{quote}
Mean FP rates for version 2:
LongBloomFilterTest: 0.997967059178744
LongLegacyBloomFilterTest: 0.997908061594203
Mean FP rates for version 3:
LongBloomFilterTest: 0.998045621980676
LongLegacyBloomFilterTest: 0.998863888888889
{quote}
The Murmur3 hash produces different results than Murmur2 so new bloomfilter
compatibility code was required. Seems as long as we don't change the default
there won't be a problem.
> add murmur3 hash
> ----------------
>
> Key: HBASE-9631
> URL: https://issues.apache.org/jira/browse/HBASE-9631
> Project: HBase
> Issue Type: New Feature
> Components: util
> Affects Versions: 0.98.0
> Reporter: Liang Xie
> Assignee: Liang Xie
> Attachments: HBase-9631.txt, HBase-9631-v2.txt
>
>
> MurmurHash3 is the successor to MurmurHash2. It comes in 3 variants - a
> 32-bit version that targets low latency for hash table use and two 128-bit
> versions for generating unique identifiers for large blocks of data, one each
> for x86 and x64 platforms.
> several open source projects have added murmur3 already, like cassandra,
> mahout, etc.
> I just port the murmur3 from MAHOUT-862. due to compatibility, let's keep the
> default Hash algo(murmur2) without changing.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
