If I were paranoid.....

then I'd find some kind of sinister conspiracy in the mysterious and utter lack of any information anywhere I can find on the internet which would provide a simple comparison ON THE SAME HARDWARE, of the leading database products all doing the same thing - for example, the TPC-C benchmarks. Of course, the reality is much simpler - no paranoid conspiracy, but instead, most vendor license agreements specifically prohibit publication of benchmark figures and no-one appears to have the courage to break ranks and actually publish the information. Does anyone out there know of any reasonably reliable source of comparative information - surely someone must have this information, and now is a good time to publish it, given that Microsoft are tied up in court and can't be their normal nasty vindictive selves until they are done with the DOJ. [I base this comment on the book "Hard Drive" in which business practices so loathsome they defy the imagination are clearly documented - I particularly dislike what MS did to prevent Mac Basic ever being released] Byte magazine, just before its untimely demise, bravely tried to do a comparison and then found it had to suppress the results as a result of vendor harassment. IBM, bless them, said it was fine to publish DB2 results. From memory I think Informix were willing to allow it too but Oracle and Microsoft said no. I am also curious about comparative benchmarks between SQL Server 6.5 and SQL Server 7. Microsoft would have us believe SQL 7 is hugely faster; other sources have suggested the performance improvements are only around 10% in many cases. I can find lots of postings asking for this information and lots of dead silence. I accept that vendors believe it's unfair to compare their products in such a simplistic way but this provides a very valuable antidote to vendor hype. It seems ludicrous that comparative benchmarks for hardware are freely available, and many software products are benchmarked comparatively, but somehow the big four database vendors appear have managed to suppress any objective comparison of their wares. Can anyone help?. If we continue to allow vendors to suppress public debate we are making serious trouble for ourselves. Can you imagine buying a new car and having GM tell you 'don't you dare write anything for an auto magazine about fuel economy and performance. Your license agreement prohibits publication of that information'. Well, when you buy an RDBMS license, that's what you signed up for - check the small print!.

Andrew, TPC benchmarks are one tiny way of comparing databases. I am 99% certain that YOUR databases won't function anything like the tpcc/d ones and so isn't the benchmark a little false? If you want to see what the differences are between dbms's for all systems then you can test them on your hardware with your systems. But there are plenty of figures on www.tpc.org. And for what it's worth don't you think that if Oracle could beat MS's top figure for tpcc on a 4-way Intel type box with NT (Compaq, HP, Dell etc.) then they'd have published this themselves - they can't so they haven't. Similarly MS can't beat Oracle/IBM on the tpc-D scores at the moment (long story there), so they haven't published any. >then I'd find some kind of sinister conspiracy in the mysterious and utter >lack of any information anywhere I can find on the internet which would >provide a simple comparison ON THE SAME HARDWARE, of the leading database >products all doing the same thing - for example, the TPC-C benchmarks. Neil Pike MVP/MCSE. Protech Computing Ltd (Please post ALL replies to the newsgroup only unless indicated otherwise) For SQL FAQ entries see www.ntfaq.com/sql.html and http://www.swynk.com/faq/sql/sqlserverfaq.asp and GO MSSQL lib 1 on Compuserve

It's real simple: The people who market a good product and do so honestly are not afraid of independant comparisons (assuming they are not cooked and are done by someone who at least knows that ACCESS is not a database.) The folks that refuse comparison have something to hide. Sadly, even if their product ran faster for your particular requirements, why would you want to do business with companies like that? My two cents. Bill Grant Andrew Mayo wrote: > then I'd find some kind of sinister conspiracy in the mysterious and utter > lack of any information anywhere I can find on the internet which would > provide a simple comparison ON THE SAME HARDWARE, of the leading database > products all doing the same thing - for example, the TPC-C benchmarks. > > Of course, the reality is much simpler - no paranoid conspiracy, but > instead, most vendor license agreements specifically prohibit publication of > benchmark figures and no-one appears to have the courage to break ranks and > actually publish the information. > > Does anyone out there know of any reasonably reliable source of comparative > information - surely someone must have this information, and now is a good > time to publish it, given that Microsoft are tied up in court and can't be > their normal nasty vindictive selves until they are done with the DOJ. [I > base this comment on the book "Hard Drive" in which business practices so > loathsome they defy the imagination are clearly documented - I particularly > dislike what MS did to prevent Mac Basic ever being released] > > Byte magazine, just before its untimely demise, bravely tried to do a > comparison and then found it had to suppress the results as a result of > vendor harassment. IBM, bless them, said it was fine to publish DB2 results. > From memory I think Informix were willing to allow it too but Oracle and > Microsoft said no. > > I am also curious about comparative benchmarks between SQL Server 6.5 and > SQL Server 7. Microsoft would have us believe SQL 7 is hugely faster; other > sources have suggested the performance improvements are only around 10% in > many cases. > > I can find lots of postings asking for this information and lots of dead > silence. I accept that vendors believe it's unfair to compare their products > in such a simplistic way but this provides a very valuable antidote to > vendor hype. It seems ludicrous that comparative benchmarks for hardware are > freely available, and many software products are benchmarked comparatively, > but somehow the big four database vendors appear have managed to suppress > any objective comparison of their wares. Can anyone help?. > > If we continue to allow vendors to suppress public debate we are making > serious trouble for ourselves. Can you imagine buying a new car and having > GM tell you 'don't you dare write anything for an auto magazine about fuel > economy and performance. Your license agreement prohibits publication of > that information'. Well, when you buy an RDBMS license, that's what you > signed up for - check the small print!.

Hi Bill, Okay, you want an independent benchmark of say Oracle Vs Sybase or Informix Vs Sybase or Oracle Vs Informix? Fine, now tell me who you want to win? I need to know so that I can tweak the benchmark. I am not joking. Just by changing the order of the table names in the where clause of a SQL join I can impact the performance of a query. If it turns out that vendor A's RDBMS uses the first table named as the driving table and vendor B's RDBMS uses the last table named as the driving table the results of an identical SQL query will prove very embarrassing to one or the other. Benchmarking is an art and no vendor wishes to have benchmark results posted by the "Paint by the numbers" level people when revenues are at stake. So, if you want to see benchmarks the Transaction Processing Council (TPC) benchmarks are about as unbiased as you can get. (A lot of people argue that the benchmarks don't reflect reality, but the TPC is the only game in town.) Contact your HARDWARE vendor and you should be able to get some fair comparisons. I will cheerfully concede that the configurations used to get the benchmark results don't reflect reality (what, you mean your 4 way NT server doesn't have 80 some odd disk drives?), every vendor has the opportunity of having their best people develop the benchmarks for their products. Based on the assumption that the best people working for any one vendor are as good at tweaking the RDBMS as the best people working for any other vendor the results offer ROUGH guidelines. regards Jerry Gitomer C. William Grant wrote in message <36D2328A.FA89CC2B@ix.netcom.com>... >It's real simple: The people who market a good product and do so honestly >are not afraid of independant comparisons (assuming they are not cooked >and are done by someone who at least knows that ACCESS is not a database.) The >folks that refuse comparison have something to hide. Sadly, even if their >product >ran faster for your particular requirements, why would you want to do business >with companies like that? > >My two cents. >Bill Grant > >Andrew Mayo wrote: > >> then I'd find some kind of sinister conspiracy in the mysterious and utter >> lack of any information anywhere I can find on the internet which would >> provide a simple comparison ON THE SAME HARDWARE, of the leading database >> products all doing the same thing - for example, the TPC-C benchmarks. >> >> Of course, the reality is much simpler - no paranoid conspiracy, but >> instead, most vendor license agreements specifically prohibit publication of >> benchmark figures and no-one appears to have the courage to break ranks and >> actually publish the information. >> >> Does anyone out there know of any reasonably reliable source of comparative >> information - surely someone must have this information, and now is a good >> time to publish it, given that Microsoft are tied up in court and can't be >> their normal nasty vindictive selves until they are done with the DOJ. [I >> base this comment on the book "Hard Drive" in which business practices so >> loathsome they defy the imagination are clearly documented - I particularly >> dislike what MS did to prevent Mac Basic ever being released] >> >> Byte magazine, just before its untimely demise, bravely tried to do a >> comparison and then found it had to suppress the results as a result of >> vendor harassment. IBM, bless them, said it was fine to publish DB2 results. >> From memory I think Informix were willing to allow it too but Oracle and >> Microsoft said no. >> >> I am also curious about comparative benchmarks between SQL Server 6.5 and >> SQL Server 7. Microsoft would have us believe SQL 7 is hugely faster; other >> sources have suggested the performance improvements are only around 10% in >> many cases. >> >> I can find lots of postings asking for this information and lots of dead >> silence. I accept that vendors believe it's unfair to compare their products >> in such a simplistic way but this provides a very valuable antidote to >> vendor hype. It seems ludicrous that comparative benchmarks for hardware are >> freely available, and many software products are benchmarked comparatively, >> but somehow the big four database vendors appear have managed to suppress >> any objective comparison of their wares. Can anyone help?. >> >> If we continue to allow vendors to suppress public debate we are making >> serious trouble for ourselves. Can you imagine buying a new car and having >> GM tell you 'don't you dare write anything for an auto magazine about fuel >> economy and performance. Your license agreement prohibits publication of >> that information'. Well, when you buy an RDBMS license, that's what you >> signed up for - check the small print!. > > >