[
https://issues.apache.org/jira/browse/HBASE-6222?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13393374#comment-13393374
]
Andrew Purtell edited comment on HBASE-6222 at 6/16/12 7:38 PM:
----------------------------------------------------------------
Thus far we have various hooks and a proposal for an out of band approach:
HBASE-2893, HBASE-4477, API object attributes (HBASE-3811, HBASE-3921,
HBASE-3931, etc.). The impetus (as I understand it) for an out of band approach
is the special nature of KV and the desire to avoid adding additional
conditionals/complexity in code wherever it touches KV. That is balanced by the
problems an out of band approach causes for minimizing overheads when tags are
present, as Jon mentions, without locality groups (a significant change) there
is extra IO. I see this as a more conservative approach that won't perturb
performance sensitive members of our community, because there is zero overhead
unless configuration is toggled on, coprocessors are loaded, and such.
However I am not arguing for or against any particular approach. The
suggestions by Jon, Lars, and Jimmy are the start of a design proposal but
should be pulled together into an actual proposal (as another JIRA?). This is
my point. Adding even only one byte to every KV has a very significant effect
on storage and IO when the data set is large. Adding a new type avoids that but
what is the impact in the code? And it wouldn't be one new type right, there
would be a tagged analogue for every mutator? A convincing exploration of these
issues would go a long way here IMHO.
was (Author: apurtell):
Thus far we have various hooks and a proposal for an out of band approach:
HBASE-2893, HBASE-4477, API object attributes (HBASE-3811, HBASE-3931, etc.).
The impetus (as I understand it) for an out of band approach is the special
nature of KV and the desire to avoid adding additional conditionals/complexity
in code wherever it touches KV. That is balanced by the problems an out of band
approach causes for minimizing overheads when tags are present, as Jon
mentions, without locality groups (a significant change) there is extra IO. I
see this as a more conservative approach that won't perturb performance
sensitive members of our community, because there is zero overhead unless
configuration is toggled on, coprocessors are loaded, and such.
However I am not arguing for or against any particular approach. The
suggestions by Jon, Lars, and Jimmy are the start of a design proposal but
should be pulled together into an actual proposal (as another JIRA?). This is
my point. Adding even only one byte to every KV has a very significant effect
on storage and IO when the data set is large. Adding a new type avoids that but
what is the impact in the code? And it wouldn't be one new type right, there
would be a tagged analogue for every mutator? A convincing exploration of these
issues would go a long way here IMHO.
> Add per-KeyValue Security, get federal funding for HBase?
> ---------------------------------------------------------
>
> Key: HBASE-6222
> URL: https://issues.apache.org/jira/browse/HBASE-6222
> Project: HBase
> Issue Type: New Feature
> Components: security
> Reporter: stack
>
> Saw an interesting article:
> http://www.fiercegovernmentit.com/story/sasc-accumulo-language-pro-open-source-say-proponents/2012-06-14
> "The Senate Armed Services Committee version of the fiscal 2013 national
> defense authorization act (S. 3254) would require DoD agencies to foreswear
> the Accumulo NoSQL database after Sept. 30, 2013, unless the DoD CIO
> certifies that there exists either no viable commercial open source database
> with security features comparable to [Accumulo] (such as the HBase or
> Cassandra databases)..."
> Not sure what a 'commercial open source database' is, and I'm not sure whats
> going on in the article, but tra-la-la'ing, if we had per-KeyValue 'security'
> like Accumulo's, we might put ourselves in the running for federal
> contributions?
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
