Re: (long) Sniffing redo logs to maintain cache consistency?

In article <Xns9331E0264F97Tokenthis_at_210.49.20.254>, Noons says...
>
>Following up on Joel Garry, 01 Mar 2003:
>
>>
>> Acceptable performance has been < 2 seconds since the 70's.
>>
>
>Let's be fair: no one needs better than <2secs. What we
>do now is process a LOT more in that timeframe and we're
>able to do it with an order of magnitude more clients.
>
>Unless we're using a Java cache, that is!
><g, d&r!>
>
>
>
>--
>Cheers
>Nuno Souto
>nsouto_at_optusnet.com.au.nospam

There ARE people out there who need this sort of response time. One system I worked on recently had a less than 2 second round trip performance requirement, and had to have an exception queue for transactions that didn't meet this. End result otherwise was a lot of packages on the floor at this shipping place, not a pretty site.

However, I agree that when most sites say they need this sort of response time, it's really not true. Will their business stop if they don't get it? Probably not. I guess I just get lucky (if you can call it that) to work on these BUS environments (BIG UGLY SYSTEMS - my new acronym for the day! ;)

HTH. Additions and corrections welcome.

Pete

SELECT standard_disclaimer, witty_remark FROM company_requirements; Received on Mon Mar 03 2003 - 11:29:34 CST