Re: efficient compare

From: Jay Dee <ais01479_at_aeneas.net>
Date: Mon, 24 Apr 2006 00:51:39 GMT
Message-ID: <viV2g.6769$YI5.5996@tornado.ohiordc.rr.com>

Andersen wrote:
> Jay Dee wrote:
>

>>  From reading the clarifications made in subsequent posts to this
>> discussion, it seems that you don't have two sets, you have one,
>> and the issue isn't how to reconcile more than one set but how to
>> maintain one set in more than one place.

>
>
> Agree,
>
>

>> Which begs the question: does it have to be more than one place?
>> is there no acceptable way to connect data sources and sinks to
>> one database?

>
>
> Yes, it needs to be in 100 places, eventual consistency is all that
> matters.
>

>> It seems you've already answered this question and are trying to
>> figure out how to maintain a distributed database.  I suggest you
>> check out the distributed databases already on the market (dismal
>> performers) or check out Stonebraker's work-in-process (tolerates
>> inconsistencies "for a while.")

>
>
> Doubt any database systems would help. I am interested in algorithms,
> not off the shelf dbms, becaues I doubt they scale to 100 or 500
> nodes... geographically distributed... It might seem like asking for a
> lot, but I am ready to accept eventual consistency.

You know, when it was just "Table A" and Table B," I was thinking, "Why not exchange tapes?" A couple tapes and FedEx is quicker than networks when lots of data are involved.

But it seems that you've got lots of data in lots of places. And, I'll bet, more than two tables, right? This is a very different situation.

And you expect to find answers on c.d.t? I admire your optimism, but I'd need some portraits of presidents before telling anyone how I did it. Received on Sun Apr 23 2006 - 19:51:39 CDT