incubator-gora-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Lewis John Mcgibbney <lewis.mcgibb...@gmail.com>
Subject Re: Gora & Riak
Date Thu, 26 Jan 2012 11:43:17 GMT
Hey Nate,

I completely understand where you coming from. I appreciate your comments
as well and this is possibly why there has been no substantiated moves to
implement this before. There are some open issues on the Jira relating to
implementing different datastores, the reason that some of them are still
open and that little prgress has been made is that they might not work in
harmony with the underlying gora-core logic, so I understand the point you
make in a similar context.

In all honesty I am happy that this discussion has morphed into the larger
issue of k/v stores instead of simply drumming up support for a gora-riak
store. I think it's important for the project to have some sort of defined
strategy moving forward... We've just received board approval for
Graduation :0)

Definitely within the ongoing scope is to improve on the "core" part of the
Gora project, which is to achieve unparalleled object mapping support for
Column stores fully embracing the Hadoop paradigm.

Thanks for now

Lewis

On Wed, Jan 25, 2012 at 4:11 PM, Nate McCall <nate@datastax.com> wrote:

> Well what interested me initially about Gora was the focus on columnar
> data stores.
>
> I've spent some time on ORM tiers and related with Apache Cassandra
> integration (hector-jpa, hector-ORM, and a JSR-347 impl), and I just
> want to throw up a warning about the level of effort required to
> engineer around impedance mismatch when you start targeting
> fundamentally different architectures.
>
> Now the differences between k/v and columnar datastores are at least
> negotiable (document stores as well to some degree), but integrating
> the number of different storage architectures into the same ORM
> plumbing as you suggest will quickly become a headache.
>
> (Full disclosure - I am currently under the employ of a commercial
> Apache Cassandra vendor).
>
>
> On Wed, Jan 25, 2012 at 2:14 PM, Lewis John Mcgibbney
> <lewis.mcgibbney@gmail.com> wrote:
> > Hi Nate,
> > Thanks for shipping in on this one.
> > Based on our vision for data persistence; we describe within the scope to
> > support object mapping to Column stores such as HBase, Cassandra,
> > Hypertable; key-value stores such as Voldermort, Redis, etc; SQL
> databases,
> > such as MySQL, HSQLDB, flat files in local file system of Hadoop HDFS.
> >
> > I assume at some stage in the future, someone will come along and
> implement
> > a module for gora-voldermort or gora-redis... so maybe gora-riak?
> >
> > I think a big part of this is to open up Gora as a project to as many
> > communities as possible, if people have problems which Gora can address
> and
> > help to solve then maybe we are missing an oppertunity if we don't act. I
> > am personally not using Riak but some guys I know are using it, I just
> > wondered if gora-riak was a go'er...?
> >
> > Thanks for now
> > Lewis
> >
> > On Tue, Jan 24, 2012 at 4:57 PM, Nate McCall <nate@datastax.com> wrote:
> >
> >> Riak is a key/value storage architecture much closer to DynamoDB. Is
> >> this a direction you want to go?
> >>
> >> Of the k/v stores out there, it is a mature product, though.
> >>
> >> On Tue, Jan 24, 2012 at 8:47 AM, Lewis John Mcgibbney
> >> <lewis.mcgibbney@gmail.com> wrote:
> >> > Hi Guys,
> >> >
> >> > I've been speaking with a friend recently who has started learning and
> >> > using Riak. I wonder if anyone else has experiences using this store
> and
> >> > whether or not a gora-riak module would be worth logging? I realize
> that
> >> > there is plenty enough work to be getting done with Gora, however
> there
> >> > might be a possibility that some students would work on a project
> such as
> >> > this.
> >> >
> >> > Thanks
> >> >
> >> > Lewis
> >> >
> >> > --
> >> > *Lewis*
> >>
> >
> >
> >
> > --
> > *Lewis*
>



-- 
*Lewis*

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message