Re: [bailey-developers] lattice master

[Ning L. <nin...@gm...>]

We can have a concept of a shard. I meant to say it'd be different from
the shard in my postings on Lucene general.

Here, we can say the distinct starting positions of all the nodes divide
the ring into shards. A node serves one or multiple continuous shards.
So, a node serves a range of docids.

Regards,
Ning

On Feb 10, 2008 12:14 PM, Yonik Seeley <yo...@ap...> wrote:

> On Feb 10, 2008 2:07 PM, Ning Li <nin...@gm...> wrote:
> > As in consistent hashing, the entire range of docids (or their hashes)
> is
> > treated
> > as a ring. There is no clear concept of a shard here.
>
> It seems to me like consistent hashing could happily live with the
> concept of a shard.
> All docids that belong to a single physical node form a shard.
> All replicas of a shard could be the same.
> I've not thought through all the implications though... I'm still
> reading the Dynamo paper myself (and after that will be looking into
> zookeeper I suppose).
>
> -Yonik
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> bailey-developers mailing list
> bai...@li...
> https://lists.sourceforge.net/lists/listinfo/bailey-developers
>