VaKo Posted June 29, 2007 Share Posted June 29, 2007 http://www.racegroups.com/reddittest.php?S...dit&Phase=2 It would appear that apache doesn't fair well... Quote Link to comment Share on other sites More sharing options...
mubix Posted June 29, 2007 Share Posted June 29, 2007 Actually, Apache is kicking ISS' arse. Let me explain: HTML was not created to dual parallel loading, so since the IIS frame is first in the HTML document it gets that fraction of a second before the Apache server gets the call. That in of it's self doesn't matter as much as since Apache gets the call second it doesn't make it's SQL queries until after IIS does. And on occasion it is STILL loading faster while using HALF the amount of ram. They are also redireting through Alexia, so this may just be a ploy to bump the sites' stats Quote Link to comment Share on other sites More sharing options...
VaKo Posted June 29, 2007 Author Share Posted June 29, 2007 At the time of postong the IIS6 page was loading instantly, while the apache page took several seconds to appear. Quote Link to comment Share on other sites More sharing options...
mubix Posted June 29, 2007 Share Posted June 29, 2007 06/29 8:50 am pacific time, cleared stats data and started with new code. 9:35 am. found php ini error on apache. increased memory, reset stats and restarted. Quote Link to comment Share on other sites More sharing options...
VaKo Posted June 29, 2007 Author Share Posted June 29, 2007 Which would make it 10am when I posted... Not long granted, but IIS still loaded faster than apache. Quote Link to comment Share on other sites More sharing options...
Sparda Posted June 29, 2007 Share Posted June 29, 2007 Also, Apache is running on Windows. Not the best option. Quote Link to comment Share on other sites More sharing options...
SomeoneE1se Posted June 30, 2007 Share Posted June 30, 2007 adding a sleep time in the page that called it could work, then they are not being call at the same time and no more "Buy A gets' to runs it's sql querys before B and that's why it's slow" Quote Link to comment Share on other sites More sharing options...
omega_ion Posted June 30, 2007 Share Posted June 30, 2007 I'm sticking with Apache just because if bugs are found, I know they will get patched much faster. Quote Link to comment Share on other sites More sharing options...
cooper Posted July 1, 2007 Share Posted July 1, 2007 I checked the link yesterday, and again today. Yesterday the load time was measured as 1.6-1.7 for each server, with IIS taking about 2-3 times more memory in the process. I look today and the times are 0.61 vs 0.64 in favor of IIS, but again while using triple the ram. When you're running a HUGE server farm this might be deemed a significant difference in response time, but for most people I doubt it would matter. Same with the memory. So Apache needs 10 while IIS needs 30. Most machines come with 1 gig by default, and probably more. It's not wise to be wasteful, but let's not get overboard, shall we? Quote Link to comment Share on other sites More sharing options...
VaKo Posted July 1, 2007 Author Share Posted July 1, 2007 I would like to see how apache on linux tuned by an apache expert fairs against IIS tuned by a IIS expert to be honest. From what I've seen when you get to HUGE sites or your being /.'d IIS does seem to hold its own against apache. IE the type of situation where cost isn't the issue. Quote Link to comment Share on other sites More sharing options...
elmer Posted July 1, 2007 Share Posted July 1, 2007 I am using the latest version of Opera with Windows XP Home on a laptop with 512 MB RAM, 1.6 GHz Turion 64, and an internet speed of around 1-3 MBits per second. Here's my load times and everything (as of July 1, 6:18 PST): IIS: Secs to load this run 0.69360303878784 Average Secs to load all runs 1.7447162742801 Memory used 27262976 Apache: Secs to load this run 0.64054203033447 Average Secs to load all runs 2.3421287151705 Memory used 11010048 Differences: IIS - Apache = Total Secs to load this run 0.0530610085 Average Secs to load all runs -0.597412441 Memory used 16252928 Quote Link to comment Share on other sites More sharing options...
Sparda Posted July 1, 2007 Share Posted July 1, 2007 I would like to see how apache on linux tuned by an apache expert fairs against IIS tuned by a IIS expert to be honest. Digg is slow. But it's quite possible that MySQL is the slow down there rather then just Apache. Perhaps the next test is MSSQL Vs MySQL Vs Oracle. Though I would be kind of worried if Oracle didn't come out on top since there database software is there main focus and they have money. Quote Link to comment Share on other sites More sharing options...
VaKo Posted July 1, 2007 Author Share Posted July 1, 2007 Anyone know if digg uses innoDB or myisam? I know innoDB chews threw resources but I'm interested in the relative merits compared to myisam. Quote Link to comment Share on other sites More sharing options...
jollyrancher82 Posted July 2, 2007 Share Posted July 2, 2007 Vako: http://www.slideshare.net/epee/mysql-2007-tech-at-digg-v3 Quote Link to comment Share on other sites More sharing options...
cooper Posted July 2, 2007 Share Posted July 2, 2007 Is the database really a factor here, since both servers are using it? The only thing I can think of favoring either setup is when they would both be trying to access the same DB simultaneously. I wouldn't expect the guy running this thing to be that stupid. Quote Link to comment Share on other sites More sharing options...
VaKo Posted July 3, 2007 Author Share Posted July 3, 2007 Vako: http://www.slideshare.net/epee/mysql-2007-tech-at-digg-v3 That was most useful, thanks. It would appear that a mixture of the two would be the most benifical, Quote Link to comment Share on other sites More sharing options...
Deags Posted July 3, 2007 Share Posted July 3, 2007 it would make sense to run iis on windows anyway? Quote Link to comment Share on other sites More sharing options...
VaKo Posted July 3, 2007 Author Share Posted July 3, 2007 OK, why don't we setup our own version of the test? IIS6 on Windows 2K3 vs Apache2 on *nix. Could even do a default configuration vs tweaking test. Quote Link to comment Share on other sites More sharing options...
cooper Posted July 4, 2007 Share Posted July 4, 2007 OK, why don't we setup our own version of the test? IIS6 on Windows 2K3 vs Apache2 on *nix. Could even do a default configuration vs tweaking test. I could do that easy. I've got a spare Mini-ITX with 2 harddisks and a gig of ram lying around. Should be peanuts to use for a comparison. But, alas, time... the final frontier. Quote Link to comment Share on other sites More sharing options...
wetelectric Posted July 5, 2007 Share Posted July 5, 2007 Comparisons like this are all well and good, but they do ignore the fact that if you use win2k3 you are locked in to microsoft tech. The licensing is insane and i'm sure it won't be long before they start charging for updates. I use win2k, win2k3 and apache daily at work. If you are stuck with using IIS, i'd just stick with win2k and be done with it. You have to tweak it like a mo fo though... "mmc /s" is your friend :) Quote Link to comment Share on other sites More sharing options...
wetelectric Posted July 6, 2007 Share Posted July 6, 2007 http://www.devindra.org/tech/2007/07/06/th...-fooled-the-web Quote Link to comment Share on other sites More sharing options...
VaKo Posted July 6, 2007 Author Share Posted July 6, 2007 Paranoid blogging at its best. Russ Jones Jul 6th 2007 Please consider my comments currently at reddit. I am Russ Jones, the creator of the test… 1. I didn’t try to hide my identity - I submitted the story with the russvirante account at Digg. - I gave my name and company in comments at Digg and Reddit - I expressly said I created it. 2. There is no link between us and the winner of the competition. - I use a random whois address and a psuedonym for whois because I dont want spam mail - Our IPs were both from Georgia, but theGoogleCache is hosted at Peer 1, who was not competing. 3. We used basic text and code to ensure the answer we wanted was the answer we got. Using mysql would have shown us which servers have the best combination of database caching and bandwidth throughput. We just wanted to know about bandwidth throughput. This was an intentional variable control. There is a lot more thorough information that discredits the entirity of this story included in the reddit post that i have linked to… http://reddit.com/info/242sd/comments/c243×6?context=5 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.