Page 1 of 6 1234 ... LastLast
Results 1 to 25 of 143
  1. #1
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971

    LiteSpeed Vs Apache Showdown

    It's my opinion that there are few recent, independent benchmarks comparing the performance of LiteSpeed and Apache on modern hardware. For that reason, I've challenged LiteSpeed to compare their product side-by-side with Apache on an 8 core intel Xeon 5520 system. More details on that available in this thread:

    http://www.webhostingtalk.com/showth...=918104&page=3

    But I need the help of the WHT community to decide how best to optimize apache. LiteSpeed will certainly know the in's and out's of their product well enough to portray it in the best possible light. I am by no means the world's foremost Apache expert, so I'm asking those of you who are familiar with it on a low-level, or have contributed to the project -- or are just veteran sysadmins -- to contribute your expertise and help me figure out how to give LiteSpeed a run for its money

    Also open for debate is what benchmarking software should be used. ab is one option that's been brought up, but I'd like to do it with at least two different tools.

    The results of the tests will be posted back in this thread.

    I'm a proponent of open source software, but I have nothing against LiteSpeed -- I'll be very impressed if it really does beat apache hands down like they say it does. May the best webserver win!
    Last edited by e-Sensibility; 01-12-2010 at 02:50 AM.
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution
      0 Not allowed!

  2. #2
    Join Date
    Jun 2006
    Location
    NYC
    Posts
    1,421
    I'll state that Apache will sadly lose based on our own internal tests but we're probably not the definition of Apache *experts*.

    We tested with mod_fcgid / PHP 5 / Worker vs LiteSpeed's LSAPI PHP5 FCGI implementation.

    LiteSpeed was faster not by leaps and bounds but was faster. The issue we had was we simply ran Apache out of resources and it croaked =/

    We had much better results when testing LiteSpeed vs nginx - whereas nginx rocked yet still didn't fully beat out LiteSpeed when it come to dynamic content.

    I'll definitely follow this thread, I'm curious to see what you guys come up with.

    I would make sure to make the test valid, be sure to use Apache 2.2 | MPM | PHP 5.3.1
    FiberPeer.Com | | REAL DDoS Protection | Cloud Hosting | VPS | Bare Metal Servers
    FiberPeer DDoS Mitigation | ethProxy Upgraded! | 14 Years Experience
    Visit us @ www.fiberpeer.com
      0 Not allowed!

  3. #3
    Join Date
    Jan 2010
    Posts
    11
    when test litespeed vs lighttpd?
      0 Not allowed!

  4. #4
    Join Date
    Oct 2007
    Posts
    3,995
    Quote Originally Posted by appvzcom View Post
    when test litespeed vs lighttpd?
    Well let's have this one step at a time.
    Aspiration Hosting | The Managed Cloud Server with Auto Failover
    High Availability Pure SSD CDN & Backups Included Magento Optimized

    NoFrillsReseller.com | Budget Reseller Hosting + 24x7 Live Chat Support
      0 Not allowed!

  5. #5
    Join Date
    May 2009
    Posts
    410
    Hello,

    I could help you by optimizing lighttpd. Drop me a PM when you decide to compare lighttpd with some other web-server.

    Thanks
    Regards
    KnownSRV.com - Fast, Quality and Secure MANAGED VPS with fast SAS drives in h/w RAID10
    Full network and hardware customization possible - let's talk your needs
    █ PayPal, Skrill(MoneyBookers), Payza(AlertPay), PerfectMoney, WebMoney and BitCoins accepted
      0 Not allowed!

  6. #6
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    Quote Originally Posted by SceneSRV View Post
    Hello,

    I could help you by optimizing lighttpd. Drop me a PM when you decide to compare lighttpd with some other web-server.

    Thanks
    Regards
    If you're willing to configure it then I don't see what harm could come of testing out lighttpd as well. Or nginx for that matter. But I'm not going to commit myself to doing that much work -- so if you want to see it tested, you have to do the configuration.

    But I'm primarily interested in seeing how apache does -- come on people, let's hear your arcane and effective optimization strategies.
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution
      0 Not allowed!

  7. #7
    Join Date
    May 2009
    Posts
    410
    Quote Originally Posted by e-Sensibility View Post
    If you're willing to configure it then I don't see what harm could come of testing out lighttpd as well. Or nginx for that matter. But I'm not going to commit myself to doing that much work -- so if you want to see it tested, you have to do the configuration.

    But I'm primarily interested in seeing how apache does -- come on people, let's hear your arcane and effective optimization strategies.
    Hey,

    I will do it if you could provide a testing server for that purpose. I was using the LSWS web-server 2 years ago and then I switched to Lighttpd and I'm pretty sure I would get better results with Lighttpd than mistwang with LiteSpeed.

    It has to be tested on the same configuration/setup as LiteSpeed and only web-server has to be optimized.

    It would be great if someone could come up with Apache or nginx setup/optimization, I would like to see the newest results and I believe everyone else on the scene would like this.

    Regards
    KnownSRV.com - Fast, Quality and Secure MANAGED VPS with fast SAS drives in h/w RAID10
    Full network and hardware customization possible - let's talk your needs
    █ PayPal, Skrill(MoneyBookers), Payza(AlertPay), PerfectMoney, WebMoney and BitCoins accepted
      0 Not allowed!

  8. #8
    I will do it if you could provide a testing server for that purpose. I was using the LSWS web-server 2 years ago and then I switched to Lighttpd and I'm pretty sure I would get better results with Lighttpd than mistwang with LiteSpeed.
    I will certainly take that challenge as well.
    Anyone would like to do it for nginx? Let's show down!

    I prediction will be LiteSpeed >> Nginx >> Lighttpd >> Apache, it is based on our past experience with those products.
    LiteSpeed Web Server by http://www.litespeedtech.com
    Best PHP and Ruby On Rails hosting platform,
    Completely Apache interchangeable
    Compatible with all hosting control panels.
      0 Not allowed!

  9. #9
    Join Date
    Oct 2008
    Posts
    533
    This is interesting, though I'm curious if anyone has a site/server busy enough for testing. Even then it would piss of a lot of site visitors when the site goes down for swapping webservers except the case of apache -> litespeed.
    Perhaps get a host that is willing to swap their heavily optimized apache + worker on a busy server with litespeed then compare the resource difference?
    Last edited by Scientist; 01-12-2010 at 03:31 PM.
      0 Not allowed!

  10. #10
    Also open for debate is what benchmarking software should be used. ab is one option that's been brought up, but I'd like to do it with at least two different tools.
    Following benchmarking tools are in my mind,

    1. ab
    2. httperf
    3. Siege
    4. curl-loader
    5. http_load

    I have experience with ab and httperf, both are single process, single thread, event driven, the performance is good, but I have to run multiple instances and add-up the results to benchmark more than one lshttpd process. in certain situation, one lshttpd process will outperform one "ab" or "httperf" process. Those tools are similar, the result is similar, so only need use one. "ab" is slightly easier to use. so, my vote is on "ab".

    I have no experience with other tools listed, but I have been interested in "Seige" for long time, this tool can be used to simulate real world load by providing a file containing a list of URL (access log file?). Not sure about its performance and whether it is multi-threaded or not. This one is the second one I would like to use.

    "curl-loader" is like "seige" which can be used to simulate real-world traffic from description on their web site, no experience, may not perform well if it uses libcurl.

    "http_load" is similar to "ab".
    LiteSpeed Web Server by http://www.litespeedtech.com
    Best PHP and Ruby On Rails hosting platform,
    Completely Apache interchangeable
    Compatible with all hosting control panels.
      0 Not allowed!

  11. #11
    Quote Originally Posted by Zsev View Post
    This is interesting, though I'm curious if anyone has a site/server busy enough for testing. Even then it would piss of a lot of site visitors when the site goes down for swapping webservers except the case of apache -> litespeed.
    Perhaps get a host that is willing to swap their heavily optimized apache + worker on a busy server with litespeed then compare the resource difference?
    e-Sensibility will offer a dual xeon 5520 server for this LiteSpeed vs Apache showdown as mentioned in another thread.
    LiteSpeed Web Server by http://www.litespeedtech.com
    Best PHP and Ruby On Rails hosting platform,
    Completely Apache interchangeable
    Compatible with all hosting control panels.
      0 Not allowed!

  12. #12
    Join Date
    Mar 2008
    Posts
    1,715
    I'm by no means an expert in optimizing Apache either, but I have a requests: please test Apache with both mpm_worker and mpm_event.
    Jamie @ Sabrienix
    Now with Mumble Hosting!
      0 Not allowed!

  13. #13
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    Mistwang and ScenSRV -- PM me with the details of when you'll be able to setup your litespeed and lighttpd respectively. We've got to get this show on the road . . . won't have this machine sitting around forever

    As for the benchmarking tool, ab and siege sound find to me.
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution
      0 Not allowed!

  14. #14
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    Also, I think that all of the configs should be posted in the end so people can draw their own conclusions about the results that are obtained -- so keep track of everything that you change.
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution
      0 Not allowed!

  15. #15
    I will have our staff member to setup LiteSpeed, I will do some fine tuning if need. just PM me the root login.

    Also, we should define what we would like to test. Small static file? larger static file? Hello world PHP script?

    How about a real-world WordPress or vBulletin with simulated real-world load using "seige"?

    And it would be great if we as well as a third party member can cross verify the test results.
    LiteSpeed Web Server by http://www.litespeedtech.com
    Best PHP and Ruby On Rails hosting platform,
    Completely Apache interchangeable
    Compatible with all hosting control panels.
      0 Not allowed!

  16. #16
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    mistwang,

    PM sent.

    And yes, I'd be interested in doing all of the tests you mentioned. I'm most intrigued by the real-world simulation.

    As for third party verification, leave the configuration on there when you're done and anyone who wants to look at your handiwork will be welcome to -- they'll just have to pm me.

    A side note -- I've made an appeal on the apache user's mailing list for configuration help, so hopefully they'll come up with some interesting ways of leveling the playing field.
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution
      0 Not allowed!

  17. #17
    I'm willing to provide a custom compiled nginx binary (or 'configure' options) as well as a config but I need to know what is going to be benchmarked in this test (fcgi/static). My vote goes to httperf or ab with another program for memory info and cpu info, also I wouldn't mind all programs to first have a trail run under perf (ingo's perf tool) to get in-depth statistics (the kernel needs to have performance counters).
      0 Not allowed!

  18. #18
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    Quote Originally Posted by nwmcsween View Post
    I'm willing to provide a custom compiled nginx binary (or 'configure' options) as well as a config but I need to know what is going to be benchmarked in this test (fcgi/static). My vote goes to httperf or ab with another program for memory info and cpu info, also I wouldn't mind all programs to first have a trail run under perf (ingo's perf tool) to get in-depth statistics (the kernel needs to have performance counters).
    So far I think mistwang has provided some good suggestions as to what content we should use.

    So:

    with ab we'll test:

    -static
    -large static
    -simple php (hello world or similar)

    And with siege we'll test

    -wordpress/joomla or similar under a "real-world" load simulation

    Perf also sounds interesting and worth a shot. Any other suggestions?
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution
      0 Not allowed!

  19. #19
    Quote Originally Posted by e-Sensibility View Post
    So far I think mistwang has provided some good suggestions as to what content we should use.

    So:

    with ab we'll test:

    -static
    -large static
    -simple php (hello world or similar)

    And with siege we'll test

    -wordpress/joomla or similar under a "real-world" load simulation

    Perf also sounds interesting and worth a shot. Any other suggestions?
    Memory and CPU usage need to be accounted for since if sever x does 20,000 r/s @ 2% cpu usage while server y does 21,000 r/s @ 40% cpu usage the efficiency (which is a must in a data center) will greatly differ.
      0 Not allowed!

  20. #20
    Join Date
    May 2009
    Posts
    410
    Hey,

    Please PM me the root login.

    Thanks
    KnownSRV.com - Fast, Quality and Secure MANAGED VPS with fast SAS drives in h/w RAID10
    Full network and hardware customization possible - let's talk your needs
    █ PayPal, Skrill(MoneyBookers), Payza(AlertPay), PerfectMoney, WebMoney and BitCoins accepted
      0 Not allowed!

  21. #21
    Quote Originally Posted by nwmcsween View Post
    Memory and CPU usage need to be accounted for since if sever x does 20,000 r/s @ 2% cpu usage while server y does 21,000 r/s @ 40% cpu usage the efficiency (which is a must in a data center) will greatly differ.
    Agree completely.

    Also, we should compare web server under similar configuration for similar feature set. For example, if .htaccess is enabled for Apache, it will slow down Apache dramatically, however, Nginx and lighttpd does not support .htaccess at all, so if we compare Apache/LiteSpeed with .htaccess enabled with lighttpd/nginx, it is not fair comparison. Same thing regarding access logging.
    LiteSpeed Web Server by http://www.litespeedtech.com
    Best PHP and Ruby On Rails hosting platform,
    Completely Apache interchangeable
    Compatible with all hosting control panels.
      0 Not allowed!

  22. #22
    Join Date
    May 2009
    Posts
    410
    I still haven't received the root login. Could someone update me regarding this, are we gonna do it or no?
    KnownSRV.com - Fast, Quality and Secure MANAGED VPS with fast SAS drives in h/w RAID10
    Full network and hardware customization possible - let's talk your needs
    █ PayPal, Skrill(MoneyBookers), Payza(AlertPay), PerfectMoney, WebMoney and BitCoins accepted
      0 Not allowed!

  23. #23
    Join Date
    Mar 2009
    Location
    /usr/bin/perl
    Posts
    971
    SceneSRV, I think we should configure the webservers one at a time. LiteSpeed can go first, and then we can all take a look at the results. Next we'll configure Apache, run the benchmarks, and everyone can look at and confirm the results.

    All of us being on there at the same time adds in too many extraneous variables.

    mistwang -- since you guys have had the login for several days now, can you update us on progress?
    Ask me about CloudCentrum (coming soon) -- The complete, turn-key cloud software solution
      0 Not allowed!

  24. #24
    Join Date
    Aug 2001
    Posts
    4,028
    This is a pretty cool project you guys have going on. Subscribing.
      0 Not allowed!

  25. #25
    Join Date
    Dec 2009
    Posts
    128
    Is it easy to upgreade to lightspeed?
      0 Not allowed!

Page 1 of 6 1234 ... LastLast

Similar Threads

  1. Apache vs Litespeed
    By Alan108 in forum Web Hosting
    Replies: 4
    Last Post: 10-08-2009, 09:54 PM
  2. LiteSpeed Vs Apache?
    By leanfarrell in forum Hosting Security and Technology
    Replies: 11
    Last Post: 10-06-2009, 03:25 PM
  3. LiteSpeed vs Apache
    By LadySDevil in forum Hosting Security and Technology
    Replies: 51
    Last Post: 07-21-2009, 12:24 AM
  4. LiteSpeed -> Apache
    By goooh in forum Hosting Security and Technology
    Replies: 6
    Last Post: 07-13-2009, 12:26 AM
  5. apache vs litespeed
    By linktome in forum Hosting Software and Control Panels
    Replies: 1
    Last Post: 07-13-2008, 08:37 PM

Related Posts from theWHIR.com

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •