Re: [Postgres-xc-general] Cluster avaibility

[Magorn <ma...@gm...>]

Hi Koichi,
I didn't search feature about big data in postgress-xc. I search something
like ORACLE RAC.
Cluste Avaibility, load balancing, auto rebalancing connection to alive
node when node is stops/break.
Big data is for BI system (very big BI system), i work with OLTP
application, my except is around OLTP.

My interset about NoSQL system, it's how it run and is it possible copy
some feature to SQL system ?
I see also about In memory Datagrid like EHCache, Infinispan or Coherence,
data always available in memory distributed on several servers.

Regards,


On Tue, Jan 10, 2012 at 3:00 AM, Koichi Suzuki <ko...@in...>wrote:

> You're running so many DBMSs.   What do you expect in XC?  As I wrote, XC
> runs as a background for Hive and Sqoop.   Do you think XC can contribute
> to this type of application?  We're now focusing in OLTP like application
> so far, although XC can be applicable for analytic applicaiton too.   My
> idea is to use SQL/MED type mechanism for big data analysis, but asking to
> have "agent" for each piece of data to help scan/join/aggregate.   This
> will eliminate the current needs to load "everything" to the database for
> analysis.   "Agent" and "SQL/MED" can integrate any data source to SQL
> statements without re-loading.
>
> Object-oriented SQL extension will be useful too, for example,
> path-expression to traverse object chain, which is described in
> "Introduction to Object-Oriented Datbases" by Won Kim.   This is one on my
> favorite because this will simplify many indirect sub-queries and will be
> more intuitive.    Anyway, I'd like to learn much more about what we need
> for big data analysis.
>
> Regards;
> ---
> Koichi Suzuki
>


-- 
Magorn