-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Konstantin,
On 10/29/12 3:16 PM, verlag.preis...@t-online.de wrote: > Hi Christopher, > > -----Original-Nachricht----- >> Von: Christopher Schultz <ch...@christopherschultz.net> Datum: >> Mon, 29 Oct 2012 03:52:36 +0100 >> >> Personally, I prefer Linux based upon its friendliness to >> developers and administrators: it's got the tools we need and >> it's easy to build additional tools if necessary. > > Well, personally, I'm a fan of windows (i.e. I wouldn't voluntarily > install any other OS than Windows on my boxes and servers ;-) ) I'm glad there are some folks out there like you that like Windows as a server platform. My experiences have been less-than-stellar, but then again, I've never been a Windows admin and actually know how to fix things in *NIX, so I'm willing to put up with more from *NIX before I get angry. I have a very short fuse when it comes to Windows, unfortunately. Long ago I had a colleague talk to me about how great Microsoft SQL Server was (actually, had become, since there were the obligatory 3 versions that weren't actually worth releasing in the first place) but with the complaint that it had one big bug: it *had* to be run on Windows. I suspect things have gotten a lot better since Windows 2003 (which I guess might have been the server version of WinXP). >> The only performance-related items of which I am aware that >> sometimes give Microsoft Windows a disadvantage are: >> >> 1. Poor uptime (due to general instability and frequent >> required-reboot OS updates) > > My personal impression is that Windows (at least the current > versions of it - Win7, Server2008R2) is one of the stablest OSes > that I have seen. The only times when my Win7/2008 crashed was when > the hardware had a defect or a faulty device driver was installed. > (I remember when I played with OpenSuse linux in a VM, where I had > to restart it every 5 minutes, as after working a bit with the > GUI/KDE it got very instable or didn't react any more - but maybe > this applies only to the GUI I used or to the OpenSuse > distribution, I'm not sure). I'd never run a GUI on a production machine but that shouldn't have required you to reboot the server all the time... maybe just make your remote desktop sessions intolerable. I'm sorry you had that experience. >> 2. Limited IP stacks on non-"server" versions > > I agree, that is what I'm also not happy with. (Currently, the > upgrade version of Win8 Pro costs only 29,99 € in germany or > $39,99 in the US - I guess the server versions will not be that > cheap :) ) Well, I guess they have to make their money somehow. I just wish they would charge extra for extra features. It seems silly to have an arbitrary limit on the number of IP connections you can have just to get people to pay for a server license. I guess the problem is that the OSS community has done such a great job getting server software (httpd, courier, MySQL, etc.) running on win32 that Microsoft can't charge anything for e.g. IIS. I still don't get it, though: Exchange ain't free and you aren't going to use that with Courier anyway. The Microsoft ecosystem ... er, encourages you to use all Microsoft products together so if you really want to run a server, you'll want IIS, Exchange, SQL Server, etc. and those (specifically Exchange and SQL Server) can have higher license fees. But limiting the IP stack is just petty. >> 3. Bizarre observations when using high-resolution (or even >> ms-res) clocks and timers... seems like you can't get more than >> about 0.1-sec resolution or so reliably -- or at least plausibly >> -- on a win32 box. > > Hmm, I think this applies for outdated versions of Windows like > WinXP, which don't support HPET timers. I remember when I wrote a > java snippet at my WinXP machine at work like this: > > > long startTime = System.nanoTime(); // do something which doesn't > take much time... long duration = System.nanoTime() - startTime; > > and then being surprised that "duration" contained a negative > value... Yeah, there was that and the fact that the positive values were never smaller than something like 200ms or so. > However, Windows versions >= 6.0 should have HPET support to allow > high precision time measurement. It seems that on my Win7 machine, > using System.nanoTime() (or the StopWatch class in .Net), I can get > a resolution of about 100 nanoseconds. I'll have to check that out. Realistically, nothing that happens that fast a single time is worth timing: if you really want to know how long something takes, you'll have to do it a few million times and take the average, which removes whatever disadvantage win32 was providing in that case. Thanks, - -chris -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.17 (Darwin) Comment: GPGTools - http://gpgtools.org Comment: Using GnuPG with Mozilla - http://www.enigmail.net/ iEYEARECAAYFAlCO8IUACgkQ9CaO5/Lv0PBBsACcDLMZsYc37Y1x8uzA8z8EU0hM /i8An1tmhrSC/THfs5YQ6lxAmk7cV/45 =p5fY -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org
