Future plans

Staņislavs Koikovs stanislavs.koikovs at gmail.com
Sat Jul 2 16:54:55 EDT 2011


David, counters would be a great addition to the Riak feature set. Of
course, it's a more like post-1.0 release feature but I think a lot of
Riak users would find them useful.

By 'offline/batch processing' I meant a better support for long
running tasks on massive amounts of data, so basically I'm very happy
with your response :)

Information on spatial indexes was very helpful. As for now I was
trying to implement a Geohash based index because of the
straightforward algorithm, but now I will definitely try other
solutions as well.

Thank you all for your help!


On Fri, Jul 1, 2011 at 6:51 PM, Jeremiah Peschka
<jeremiah.peschka at gmail.com> wrote:
> For what it's worth, SQL Server uses a combination of Hilbert curves and a 4
> level quad tree. If you want to implement that sort of thing, it's pretty
> well documented both in that article Francisco shared as well as quite a few
> other places. It seems to work very well for them. I've had a few clients
> switch from Oracle to SQL Server just to get the faster spatial data types.
>
> ---
> Jeremiah Peschka
> Founder, Brent Ozar PLF, LLC
>
> On Friday, July 1, 2011 at 8:02 AM, francisco treacy wrote:
>
> Stumbled on it the other day:
>
> http://blog.notdot.net/2009/11/Damn-Cool-Algorithms-Spatial-indexing-with-Quadtrees-and-Hilbert-Curves
>
>
> 2011/7/1 Sean Cribbs <sean at basho.com>:
>
> If you can't wait for native geo-spatial indexes, there are also numerous
> techniques for creating them over distributed hash tables (like Riak).  In
> some limited use-cases, you could create flat indexes directly in Riak,
> where the keys are named for the distance along a space-filling curve (like
> Hilbert, Lebesgue/Z-curve or Peano). The flat-index technique plays nicely
> with Riak's existing MapReduce query system.
> On Fri, Jul 1, 2011 at 10:30 AM, David Smith <dizzyd at basho.com> wrote:
>
> Hi Stanislavs,
>
> On Thu, Jun 30, 2011 at 5:15 PM, Staņislavs Koikovs
> <stanislavs.koikovs at gmail.com> wrote:
>
> * Distributed counters;
> * Geospatial indexing support in search;
> * Offline / batch processing for data.
>
> As you noted, we're sorta busy with the existing features for the next
> major release. :) Insofar as counters go, we've had some discussion
> about how to do them and may take a look at it after the release -- a
> lot depends on customer needs.
>
> Can you elaborate on what you mean by offline/batch processing? We're
> working on improving the usefulness of MapReduce w/ very large jobs,
> but I'm not sure if that aligns with what you're asking about.
>
> Thanks,
>
> D.
>
> --
> Dave Smith
> Director, Engineering
> Basho Technologies, Inc.
> dizzyd at basho.com
>
> _______________________________________________
> riak-users mailing list
> riak-users at lists.basho.com
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
>
>
>
> --
> Sean Cribbs <sean at basho.com>
> Developer Advocate
> Basho Technologies, Inc.
> http://www.basho.com/
>
> _______________________________________________
> riak-users mailing list
> riak-users at lists.basho.com
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
>
> _______________________________________________
> riak-users mailing list
> riak-users at lists.basho.com
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
>
>
> _______________________________________________
> riak-users mailing list
> riak-users at lists.basho.com
> http://lists.basho.com/mailman/listinfo/riak-users_lists.basho.com
>
>




More information about the riak-users mailing list