Re: DB2 / SQL Replicator ?? <<<<<<<

no need for numbers really.. I just wanted to point out that the network is typically the bottleneck and that DRDA isn't too bad at block fetching etc. to make good use of the bandwidth. and yes, there are faster ways to do things that pure SQL but this thread was purely DRDA vs your move re the wire. I think your stuff does some compression and that costs something too. VTAM can compress also so it's hard to get numbers for all this.

In article <4cs1nr$56g_at_dub-news-svc-5.compuserve.com>, kallm_at_platinum.com wrote:

> kbradley_at_us.oracle.com (Kirk Bradley) wrote:
>
>
> >Since DRDA is based on APPC why do you say your stuff must have
> >better performance? got any numbers?
>
> Since perfomance varies widely depending on network setup and traffic,
> number are hard to give.
> Another point would be to look at total times (extract / transport /
> apply).
> Since we *can* use high performance utilities for each of these phases
> (depending on circumstances) we can be a lot faster than sql.
> Example: Converting to the right numeric format for the atrget
> database during extract on MVS, building page formats on the
> mainframe, thus eliminating any load time, etc.
>
> Main point is to go with bulk processing instead of row processing,
> where this is possible...
>
> I can try to get numbers if you give me info / parameters about
> concrete environment. My feeling is that no one can be truly
> interested in well tuned marketing numbers (" there are lies, damned
> lies and performance benchmarks... ;-)")
>
>
> Regards, Martin
> -----------------------------------------------
> kallm_at_platinum.com or 100732.511_at_compuserve.com
> http://www.platinum.com
> Opinions are mine own, may be not PLATINUMs
Received on Mon Jan 08 1996 - 00:00:00 CET