Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
HBase, mail # dev - Notes from HBase Types/Serialization Discussion


Copy link to this message
-
Re: Notes from HBase Types/Serialization Discussion
Andrew Purtell 2013-03-08, 01:33
> Stack proposed the novel idea of all of us agreeing to something and then
actually doing it, and there was consensus that this was a good idea.   :-)

LOL

Discussion was great. Thanks for sharing the notes.

On Fri, Mar 8, 2013 at 6:16 AM, Stack <[EMAIL PROTECTED]> wrote:

> (I'm sending this for Doug Meil; he is having difficulty getting this up on
> the dev list)
>
>
> Hi Devs-
>
> Notes from an HBase client conference call a few of us had yesterday.
>  Attendees:  Enis, Nick D., James Taylor, Stack, LarsH, Doug Meil and Andy
> Johnson.
>
>    - Discussion of Serialization Scheme Independent of Java
>       - After a short discussion, this is a very interesting idea but off
>       the table for the next 4-6 releases.
>       - E.g., would we have to invent serialization for C++ or Python
>       clients?  It's not a small undertaking.
>       - Nobody was aware of any other open source libraries that did this.
>    - Component Priority
>       - After a healthy discussion, the following priority emerged:
>    - 1) Serialized Types
>          - But focusing on Java.  Using some combination of Orderly and
>          Phoenix
>          - Nick/James to lead this.
>          - New Jira will be created for this effort, with sub-tasks for
>          follow-on tasks.
>       - 2) Utilities
>          - E.g., something akin to what HBASE-7221 was trying to achieve.
>           This might also be a documentation exercise on how to
> properly use Orderly
>          - I.e., utilities to build rowkeys (et al.) in code for use by
>          applications (I.e., not Hbase itself)
>          - Consensus on support for both FixedLength and VariableLength
>          rowkeys, to allow for faster scans especially in the former case.
>  But
>          still providing utilities for variable length keys because
> folks still use
>          things like URLs in rowkeys.
>       - 3) Schema
>          - Schema definitions of rowkeys and tables that could be
>          externalized to Hive, et al.
>       - HBase-Centric API
>       - With respect to Utilities and Schemas, there was consensus that
>       completely enveloping HTable is not the suggested path, but rather we
>       should look for opportunities to improve the Get, Put, Scan
> interfaces with
>       Htable but make it less tedious with byte-array handling.
>    - HBASE-7221
>       - This will be closed as "won't fixed", but will be linked from
>       whatever the new Jira is as a related ticket.
>    - Agreeing To Something And Actually Doing It
>       - Stack proposed the novel idea of all of us agreeing to something
>       and then actually doing it, and there was consensus that this was a
> good
>       idea.   :-)
>

--
Best regards,

   - Andy

Problems worthy of attack prove their worth by hitting back. - Piet Hein
(via Tom White)