Re: [bailey-developers] lattice master

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

On Fri, Feb 29, 2008 at 4:24 PM, Doug Cutting <cu...@ap...> wrote:
>  Finally, as you observed, there would be hiccups whenever a node fails.
>   Hiccups affect a small percentage of BigTable clients, only those
>  touching the tablet on the failed node.  But, in distributed search,
>  every query touches a large portion of the nodes.  So, in a 1000 node
>  cluster, a failure might delay .1% of BigTable users, but might delay
>  33% of distributed search users (assuming 3-way replication).  So search
>  can be much more sensitive to this.
>
>  So I'm not convinced that the BigTable model is as appropriate for
>  distributed full-text search as consistent hashing.  Thoughts?

I used BigTable as an example for single-write-replica. I didn't intend
to use BigTable's single-read/write-replica model for distributed full-text
search. I was thinking more like a single-write-multi-read-replica model
(using consistent hashing). I considered your proposal as a
multi-write/read-replica model.

As I said before, the multi-write/read-replica model is more powerful
than the single-write-multi-read-replica model. I was just worried about
its complexity. But now I think it can be achieved without being too
complicated. :)

Ning