Re: Looking for performance and features of databases

Al Kosteniuk <kosteniuk@sk.sympatico.ca> wrote: -Regarding DB2 blowing away Oracle (I've used both), I'd suggest that the -results are not conclusive until something more is known about the -platforms involved. Same platform: WinNT 4.0 spk1, Intel 266-MHz, 178 MB RAM, IDE drive -Personal Oracle is known to be a bit of a hog. An enormous pig ... -But -since I have been away from DB2 for so long, is there now a desktop -version of this? 5.x works fine on NT -Further, why on earth would benchmarks for Personal Oracle be considered -when it sounds like a serious networked solution is required? If anything, Personal ORACLE should be quicker (smaller footprint) than a larger version. - -Just wondered .... -ak (What was a scathing commentary on object technology is that neither Objectivity nor Versant could pass the test: both use the lame Dharma SQL engine which can not process native SQL triggers, meaning object databases can not in fact compete with RDBMS technology.)

anonymous@comp.databases.object wrote: > -Further, why on earth would benchmarks for Personal Oracle be considered > -when it sounds like a serious networked solution is required? > > If anything, Personal ORACLE should be quicker (smaller footprint) than a larger > version. Maybe but is my impression that *generally* there is a trade off made between space and speed. Therefore a smaller footprint would most likely imply poorer performance. Simon.

Dear anonymous, appreciated the thoughts in your posting. Wouldn't mind getting a direct mail from you. ak

Simon Harris wrote: > anonymous@comp.databases.object wrote: > > > -Further, why on earth would benchmarks for Personal Oracle be considered > > -when it sounds like a serious networked solution is required? > > > > If anything, Personal ORACLE should be quicker (smaller footprint) than a larger > > version. > > Maybe but is my impression that *generally* there is a trade off made between space > and speed. Therefore a smaller footprint would most likely imply poorer performance. > > Simon. Guys, Canadian & American, I have a slightly general question that I hope one or both of you could help me with. My question is related to the concept of concurrent runtime database licenses. If an organization has a growing database, say they're using something like Informix Dynamic Server, for various things entered by users through their website - also growing (in traffic level). So, the database is updated via the webserver (say, Apache). My question is, how many concurrent user licences do you suspect would be needed for certain levels of traffic. (i.e., X concurrent user licenses for 100 db accesses a day, 1000 db accesses a day, ect). I guess what I'm asking is really a question of probability & statistics (and chaos theory for that matter): 24 hrs in a day, X # db accesses a day (avg) + some math = probability of Y # of concurrent accesses to occur. I would think that those in the db industry would have general benchmarks to go by. From your experience with databases, hope you guys could give me a helpful estimate -- I really have no way hypothesizing... Thanks mucho. -Steve

Steven Parker wrote: > > Simon Harris wrote: > > > anonymous@comp.databases.object wrote: > > [SNIP] > If an organization has a growing database, say they're using something like Informix > Dynamic Server, for various things entered by users through their website - also > growing (in traffic level). So, the database is updated via the webserver (say, > Apache). My question is, how many concurrent user licences do you suspect would be > needed for certain levels of traffic. (i.e., X concurrent user [SNIP] This is such a hot topic that I am certain that Informix is in the process of developing new pricing models for WEB apps. I have no good answer. Art S. Kagel