You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by Shimi K <sh...@gmail.com> on 2008/11/26 08:26:40 UTC

Switching to HBase from HDFS

I have a system which uses HDFS to store files on multiple nodes. On
each HDFS node machine I have another application which reads the
local files. Until know my system worked only with files, HDFS seemed
like the right solution and everything worked fine. Now I need to save
additional information for every file. I thought that I might create a
central database and in this database I will create a table which will
map file name with the new data. I don't think that this is a good
solution since I will need to query this new data for each file. I
thought that since HBase is built on top of HDFS it might be better to
use it instead of a database. With HBase I will have each file
together with the new data locally on each node. I can read each file
together with any additional information.

Since I never used HBase I want to ask the community if HBase is the
right solution for my case?

--Shimi

Re: Switching to HBase from HDFS

Posted by Yi-Kai Tsai <yi...@yahoo-inc.com>.
Hi Shimi

HBase (or BigTable) is a sparse, distributed, persistent 
multidimensional sorted map ,
Jim R. Wilson have a excellent article for understanding it :
http://jimbojw.com/wiki/index.php?title=Understanding_HBase_and_BigTable


> I have a system which uses HDFS to store files on multiple nodes. On
> each HDFS node machine I have another application which reads the
> local files. Until know my system worked only with files, HDFS seemed
> like the right solution and everything worked fine. Now I need to save
> additional information for every file. I thought that I might create a
> central database and in this database I will create a table which will
> map file name with the new data. I don't think that this is a good
> solution since I will need to query this new data for each file. I
> thought that since HBase is built on top of HDFS it might be better to
> use it instead of a database. With HBase I will have each file
> together with the new data locally on each node. I can read each file
> together with any additional information.
>
> Since I never used HBase I want to ask the community if HBase is the
> right solution for my case?
>
> --Shimi
>   


-- 
Yi-Kai Tsai (cuma) <yi...@yahoo-inc.com>, Asia Regional Search Engineering.