[Orca-dev] Orca on rails?

[Cockcroft, Adrian]

Thanks Cynthia,

I was thinking of using RoR to index the location of the raw orca
datafiles on a node, then to have the service provide a choice of
protocols for moving the datafiles or extracts from them to a central
point for display by the orca/rrdtool system, and to allow for a web
based drill down into the raw data on an individual node.

If RoR is only storing an index and some derived summary info then it
will grow slowly, and it could include the location of the data as being
local, remote or archived.

Adrian

-----Original Message-----
From: Cynthia Kiser [mailto:cnk at caltech.edu] 
Sent: Thursday, May 04, 2006 10:33 AM
To: Cockcroft, Adrian
Cc: orca-dev at orcaware.com
Subject: Re: [Orca-dev] Orca on rails?

Quoting Cockcroft, Adrian <acockcroft at ebay.com>:
> 
> I'm thinking of using the Ruby on Rails framework to build a simple
and flexible web service that indexes and returns orca (and related)
data from a node. Has anyone else got any advice or experience in this
area?
> 
> Cheers Adrian

I am using orca and starting to do some Ruby on Rails development, so
your idea Sounds interesting to me. One thing that is a mismatch
between the current orca set up and rails is the datastore: a sql
database vs rrdtool. What are your intentions about that? Writing a
version of orca that stored all the data in a database doesn't sound
hard, but in addition to the issue that that locks you into an ever
increasing data store (as opposed to constant size with a round robin
database), you would have to either write SQL that was able to segment
the data and average it or you would need to have additional, derived
tables that contained averages. The later would be more efficient -
but leads to even larger database size.

-- 
Cynthia Kiser
cnk at caltech.edu