[Noisebridge-discuss] [crash] Databases

[Matt Pinner]

>From the comparison you posted I'd use HBase for two reasons:
 - Map/reduce with Hadoop
 - No single point of failure
--matt




On Tue, Aug 9, 2011 at 1:16 PM, R. Kevin Nelson
<r.kevin.nelson at gmail.com> wrote:
> Crashers!!  Bridgers!!
> Sorry for the doublepost, I know some of you read both of these lists.
>
> I'm working on some things with SafeCast and am weighing different database
> options for various prototyping things.  Mostly, I'm looking at creating a
> database server to run our datalogging that's separate from or main page.
> I'm leaning towards MongoDB, but the discussion
> at http://news.ycombinator.com/item?id=2052852 has prompted me to put a
> little more care into the decision and I thought reaching out to you all
> might help quite a bit (this is my first large web architecture project).
> Basically, we need a server that handles incoming data from radiation
> measurement devices.  The data needs to go into a database that our website
> will use to display maps.
> We're already doing this at www.safecast.org (check it out!), but we need to
> make it better.
>
> So, some characteristics we're working with:
> - for the most part, once data is logged, it doesn't need to change, but now
> and then we might have to modify data based on calibration (or something..
> why this would happen is still a little vague to me, but it came up in
> previous discussions)
> - we'll probably have some cron-job type scripts running to cache data from
> specific drives (like, in a car) so that we don't have to hit the db every
> time someone access a drive map that hasn't changed in 3 months.
> - in addition to displaying data from drives, we want to be able to display
> all data for a particular area so users can use their current location to
> find all measurements made in their area.
> - probably going to use nodeJS for the server, though I could probably be
> convinced to use python.
>
> So, what can you suggest?
> Thanks!
> -kevin
>
>