[
https://issues.apache.org/jira/browse/HBASE-22005?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16787473#comment-16787473
]
Duo Zhang commented on HBASE-22005:
-----------------------------------
And [[email protected]], the point here is not about streaming API, we just
want to use (Direct)ByteBuffer instead of byte[] so we can avoid copying data
on heap.
And for the high speed new hardware(NVMe, 3DXPoint), it is mmapped, and usually
we will expose a MappedByteBuffer in the java API, which is already enough for
us here I believe? And for high latency object storage, such as S3, I do not
see any difference between passing a <offset, len, byte[][]> and <offset, len,
ByteBuffer[]>?
Anyway, we need the StreamCapabilities API to query whether a given stream has
the ability, sadly it is only provided in hadoop-2.9+ and we still need to
support 2.7...
> Use ByteBuff's refcnt to track the life cycle of data block
> -----------------------------------------------------------
>
> Key: HBASE-22005
> URL: https://issues.apache.org/jira/browse/HBASE-22005
> Project: HBase
> Issue Type: Sub-task
> Reporter: Zheng Hu
> Assignee: Zheng Hu
> Priority: Major
> Attachments: HBASE-22005.HBASE-21879.v1.patch
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
