Re: Best database for web backend on Linux?

Strange how no-one seems to mention Informix in this debate... :-) On 22 Oct 98, at 17:52, Evan Carew wrote: > I'm not shure what YMMV means but I think I get the gist if it. I know Your Mileage May Vary. > what you mean, after having arguments about the merrits of using row level > locks (especially on a platform that only supports page locking), a switch > to a platform that does support RLLs seems like a substantial performance > improvement. Having seen this, most managers get this strange idea that > due to this, RLL capable platforms are better performers. <sigh> In the Informix world I have not seen a degredation in performance when moving to RLL from PLL. I'm sure there *is* a small (or maybe not so small) performance hit, but it's never been sufficient to become an issue. Because most of my applications have been host- centric using 4GL, I am kind of used to working at the row, rather than the set, level. And people bitch far quicker about concurrency than performance! > The only thing that is capable of changing their minds after that is a > "midnight engineering" effort that removes dependency on RLL for > comparison. Depending on the application a 2X (or greater) improvement of > both performance and capacity (on any DBMS) is not uncommon. Duly noted. However, as I said before, people will scream about inappropriate locks held long before they say "I believe we can reduce our sub-second response even further." :-) -- Ciao, Billy /Group Managing Director, The West Solutions Group: http://www.west.co.za \\ Drivel @ http://www.west.co.za/tasteless/ / \\ "Granted, Mr Wheeler's ideas are stupid and unreasonable, but he does own the company and I / think we should go along with him..." \\________________________________________________________________________________________________