Re: Examples of Highest TPS rate for "Business Transactions" (fwd)
Posted in 1996
[who said what, when removed, intentionally...'cause it doesn't matter] > Hi, > > Off late I am observing this 'informix-list' mails and I think > most of the people just think it as a trash can and dump whatever > they like (or comes into their mind). > > We all here are professionals. I request you to reply in a professional > manner. If you are not sure about the answer or you do not know then > PLEASE do not respond. > > Take the following example, What DON requested in the mail what the reply > he got ? > > ONCE AGAIN I REQUEST YOU ALL PLEASE DO NOT FILL THIS DATABASE WITH JUNK. > > Thanks & Regards, > ****************************************************** > ***** This is the DON's request and the below is the reply. > > } > I am interested in obtaining any information relating to actual customer > } > transaction rates for business transactions. I would like to see actual > } > achieved numbers instead of looking at TPC-A,B,C,D transaction ratings. This > } > is an important request as we are considering the use of Informix for a > } > highly used system. This system may have to sustain a transaction rate of > } > 100+ TPS for business related transactions. Please respond with "real" > } > numbers. > } > } On Tue, 17 Sep 1996, Billy Wheeler wrote: > } > } 17, 243, 555, 987, 223, 85, 9.75. > } > } Sorry, Don, but this *has* to be the most arbitrary request I've ever seen > } here. What kind of business are you in? What kit are you using? What do your > } transactions look like? Give us a hint, at least... > } > } *I've* never run into hassles with performance under Informix. But I have *no* > } idea of how to relate that to your requirements. > } I almost hate to get involved in this one because I KNOW I'm opening myself up to flames, but, I couldn't sit still on this one. Here goes: Billy's response, while it included a (weak?) attempt at humor (sorry, Billy), was not "junk". He did have a good point. Benchmark numbers are like statistics: they're meaningless without a context. Don did need to supply more information, because the TYPES of transactions he's running will have a great effect on the results he'd see from a benchmark. In fact, any benchmarking tests he runs will need to be geared toward the types of transactions his business expects. For example, benchmarks on I/O intensive transactions are going to be different from those that do less I/O but more CPU-intensive work. And, some systems (including hardware) do better in some environments than others. So, perhaps Billy could have been less glib in his repsonse, but I don't think it was trash-can fodder, and he did have a pretty good point underneath it all. Folks, true, we're all professionals, and we all have experienced stupid posts/responses, etc. But, part of exhibiting professionalism is showing a measure of tolerance, too. We've all done/said things where we wish we hadn't pressed that "send" key so quickly...and I'll be one of the first people that rolls my eyes at that waste of bandwidth on this, and lots of other newsgroups and mailing lists. If you have a beef with someone, it might also be more appropriate to take it to the individual, rather than air it in public... OK...'nuff said. Bash away. - Corinne -- Corinne Gregory, Consulting Services Manager ASIX, Inc. 777 108th Ave. NE, Suite 1830, Bellevue WA 98004-5118 corinne.gregory@asix.com