Results 1 to 17 of 17
  1. #1
    Join Date
    May 2006
    Posts
    1,398

    Post nginx vs litespeed benchmarks

    If anyone has a litespeed server that either throttling limits are disabled or you can whitelist the ab test server get with me to do some benchmarks. In my first bench i simply tested whmcs status page on a host I knew was running litespeed and compared to the company I worked for. So I think his litespeed throttled connections or something. But here is initial results.

    If anyone has a litespeed server and wants to cooperate with me to put up same test file and so on, dynamic and static that would be great.

    My company's nginx server
    Code:
    Server Software:        nginx
    Server Hostname:        hostingrq.com
    Server Port:            80
    
    Document Path:          /clients/status/index.php
    Document Length:        206 bytes
    
    Concurrency Level:      10
    Time taken for tests:   1.893071 seconds
    Complete requests:      100
    Failed requests:        13
       (Connect: 0, Length: 13, Exceptions: 0)
    Write errors:           0
    Non-2xx responses:      87
    Total transferred:      36543 bytes
    HTML transferred:       19456 bytes
    Requests per second:    52.82 [#/sec] (mean)
    Time per request:       189.307 [ms] (mean)
    Time per request:       18.931 [ms] (mean, across all concurrent requests)
    Transfer rate:          18.49 [Kbytes/sec] received
    
    Connection Times (ms)
                  min  mean[+/-sd] median   max
    Connect:       42   42   0.4     42      43
    Processing:    42  119 262.1     43    1591
    Waiting:       42  118 262.2     43    1591
    Total:         85  161 262.1     85    1634
    
    Percentage of the requests served within a certain time (ms)
      50%     85
      66%     86
      75%     86
      80%     86
      90%    329
      95%    801
      98%   1353
      99%   1634
     100%   1634 (longest request)
    Other host running litespeed

    Code:
    Server Software:        LiteSpeed
    Server Hostname:        ******.com
    Server Port:            80
    
    Document Path:          /clients/status/index.php
    Document Length:        128 bytes
    
    Concurrency Level:      10
    Time taken for tests:   49.688669 seconds
    Complete requests:      100
    Failed requests:        0
    Write errors:           0
    Total transferred:      29300 bytes
    HTML transferred:       12800 bytes
    Requests per second:    2.01 [#/sec] (mean)
    Time per request:       4968.867 [ms] (mean)
    Time per request:       496.887 [ms] (mean, across all concurrent requests)
    Transfer rate:          0.56 [Kbytes/sec] received
    
    Connection Times (ms)
                  min  mean[+/-sd] median   max
    Connect:       81  236 661.9     81    3108
    Processing:    94 4526 9238.6   1070   39088
    Waiting:       92 4525 9238.6   1069   39087
    Total:        177 4763 9201.6   1162   39171
    
    Percentage of the requests served within a certain time (ms)
      50%   1162
      66%   2031
      75%   4173
      80%   4255
      90%  12272
      95%  35190
      98%  38203
      99%  39171
     100%  39171 (longest request)
    need some more examples and someone with litespeed who wants to work with me on complete benchmarks

  2. #2
    Theres already been tests posted on the forums, litespeed won

  3. #3
    Join Date
    May 2006
    Posts
    1,398
    Quote Originally Posted by nwmcsween View Post
    Theres already been tests posted on the forums, litespeed won
    I don't see any benches in the last year, actually couldn't find a benchmark thread at all so if there is one it is outdated or if recent one lemme see pls

  4. #4
    Join Date
    Feb 2008
    Location
    Houston, Texas, USA
    Posts
    2,955
    Non-2xx responses: 87
    That's not good. Are you sure you have php configured properly? It looks like you ran out of php processes.

    By the way, are you looking to switch to Litespeed? Why not get a free trial and run a complete benchmark?

    Regards
    Joe / UNIXY
    UNIXy - Fully Managed Servers and Clusters - Established in 2006
    [ cPanel Varnish Nginx Plugin ] - Enhance LiteSpeed and Apache Performance
    www.unixy.net - Los Angeles | Houston | Atlanta | Rotterdam
    Love to help pro bono (time permitting). joe > unixy.net

  5. #5
    Join Date
    May 2006
    Posts
    1,398
    Quote Originally Posted by UNIXy View Post
    That's not good. Are you sure you have php configured properly? It looks like you ran out of php processes.

    By the way, are you looking to switch to Litespeed? Why not get a free trial and run a complete benchmark?

    Regards
    Joe / UNIXY
    no man I have ran litespeed and championed it for around 3 years now. I am looking for alternatives. I had turned on a few big site clients to them and used them as well, had some support issues but what really scares me about it is security, bad track record and only fixes I have seen so far as request filters, no hard code fix. And plus an open source platform for everything is ideal.

    And yes php is configured ok, its suphp and server is on cloudlinux. Was thinking of trying fcgi. But probably need to adjust some throttling and other settings on my end as well to make a true benchmark. That is what I'm trying to do - find someone with litespeed who would wanna hook up and run some benches to both, making sure both environments are capable and not throttled.

    And yeah was throttling on my part. Will fix for next bench

    new one
    Code:
    Server Software:        nginx
    Server Hostname:        hostingrq.com
    Server Port:            80
    
    Document Path:          /clients/status/index.php
    Document Length:        118 bytes
    
    Concurrency Level:      10
    Time taken for tests:   11.483002 seconds
    Complete requests:      100
    Failed requests:        0
    Write errors:           0
    Total transferred:      28800 bytes
    HTML transferred:       11800 bytes
    Requests per second:    8.71 [#/sec] (mean)
    Time per request:       1148.300 [ms] (mean)
    Time per request:       114.830 [ms] (mean, across all concurrent requests)
    Transfer rate:          2.44 [Kbytes/sec] received
    
    Connection Times (ms)
                  min  mean[+/-sd] median   max
    Connect:       42   42   0.2     42      43
    Processing:   149 1100 1125.6    714    5800
    Waiting:      148 1099 1125.5    713    5799
    Total:        191 1142 1125.6    756    5842
    
    Percentage of the requests served within a certain time (ms)
      50%    756
      66%   1353
      75%   1556
      80%   1858
      90%   2322
      95%   3279
      98%   5795
      99%   5842
     100%   5842 (longest request)
    Last edited by jon-f; 02-19-2011 at 10:17 PM.

  6. #6
    Join Date
    Feb 2008
    Location
    Houston, Texas, USA
    Posts
    2,955
    Quote Originally Posted by jon-f View Post

    And yes php is configured ok, its suphp and server is on cloudlinux.
    CloudLinux's goal is not really performance but fair share of resources. Perhaps it's throttling php (I understand you're suspecting LSWS is throttling connections). Anyway, sorry I don't have a non-production LSWS to share. Hopefully someone will.

    Regards
    UNIXy - Fully Managed Servers and Clusters - Established in 2006
    [ cPanel Varnish Nginx Plugin ] - Enhance LiteSpeed and Apache Performance
    www.unixy.net - Los Angeles | Houston | Atlanta | Rotterdam
    Love to help pro bono (time permitting). joe > unixy.net

  7. #7
    Join Date
    Mar 2003
    Location
    California USA
    Posts
    13,290
    Cloudlinux is very aggressive at throttling litespeed.

    Jon for comparison install fastcgi + apache with http://www.cloudlinux.com/docs/fastcgilve.php.

    Its not uncommon to see triple the requests a second on the fastcgi then litespeed with lsphp when using cloudlinux. I just ran into this with a customer. There also seems to be quite a delay in initiating a LVE through cloudlinux.

    For what its worth - there is a performance drop using the cloudlinux kernel over the stock centos 5 kernels. If your going for pure performance get rid of cloudlinux.

    Litespeeds cloudlinux support seems kind of flaky to me. There has also been issues where litespeed will suck up 100% cpu which has been seen on several servers by me and is documented on the cloudlinux site.

    Cloudlinux and apache seems quite good.
    Last edited by Steven; 02-19-2011 at 10:56 PM.
    Steven Ciaburri | Industry's Best Server Management - Rack911.com
    Software Auditing - 400+ Vulnerabilities Found - Quote @ https://www.RACK911Labs.com
    Fully Managed Dedicated Servers (Las Vegas, New York City, & Amsterdam) (AS62710)
    FreeBSD & Linux Server Management, Security Auditing, Server Optimization, PCI Compliance

  8. #8
    Quote Originally Posted by jon-f View Post
    I don't see any benches in the last year, actually couldn't find a benchmark thread at all so if there is one it is outdated or if recent one lemme see pls
    apache vs. nginx vs. litespeed vs. lighttpd

  9. #9
    Join Date
    May 2006
    Posts
    1,398
    Quote Originally Posted by nwmcsween View Post
    man please quit sig spamming this thread if you have nothing useful to offer. AGAIN, no benchmarks within a year and only benchmarks on that thread are from the maker of litespeed. Face it, you tried to jump in the thread with a "hey why dont you use the search results" when you didnt even fully read my thread in the first place nor seen updates/replies to see what the goal is here.
    I hate to be rude, but really though, why even jump in this thread unless you have either a nginx or lsws server to benchmark.

    PLUS the goal is to compile new data here, NOT to search for old data and old threads about vs. vs vs.

    For others who didnt bother to read, I am looking for someone to cooperate and provide new benchmarks with. To compile new data on this

  10. #10
    Join Date
    May 2006
    Posts
    1,398
    Quote Originally Posted by Steven View Post
    Cloudlinux is very aggressive at throttling litespeed.

    Jon for comparison install fastcgi + apache with http://www.cloudlinux.com/docs/fastcgilve.php.

    Its not uncommon to see triple the requests a second on the fastcgi then litespeed with lsphp when using cloudlinux. I just ran into this with a customer. There also seems to be quite a delay in initiating a LVE through cloudlinux.

    For what its worth - there is a performance drop using the cloudlinux kernel over the stock centos 5 kernels. If your going for pure performance get rid of cloudlinux.

    Litespeeds cloudlinux support seems kind of flaky to me. There has also been issues where litespeed will suck up 100% cpu which has been seen on several servers by me and is documented on the cloudlinux site.

    Cloudlinux and apache seems quite good.
    Cloudlinux seems to be pretty good for shared hosting and ddos protected hosting but what is the equivalent or another way to limit per user resources?

    I tried to install fastcgi on the httpd backend a time or two but all went to hell. I need a test server for that. The production server I tried on resulted in some downtime, Ill have to find the error I was getting but i may try it again soon.

    Also on another note I think all hosts and server users should try to use open source software anyway instead of payware like litespeed. I don't think most people realize how many security issues litespeed has went through. Each of which I have seen only fixed by adding a mod security type request filter. It is still vulnerable to null byte, buffer overflows and other methods via bypass of the filters. I have seen it shown to me
    Last edited by jon-f; 02-20-2011 at 12:40 AM.

  11. #11
    Quote Originally Posted by jon-f View Post
    man please quit sig spamming this thread if you have nothing useful to offer...
    Considering I ran the benchmarks for nginx on that thread here and absolutely nothing has changed in the way nginx handles connections as per me reading the change logs due to having to maintain distro *packages* I think I may know what I'm talking about.

  12. #12
    Join Date
    May 2006
    Posts
    1,398
    Quote Originally Posted by nwmcsween View Post
    Considering I ran the benchmarks for nginx on that thread here and absolutely nothing has changed in the way nginx handles connections as per me reading the change logs due to having to maintain distro *packages* I think I may know what I'm talking about.
    Well bench it again if you want. Hopefully someone will bench lsws too

  13. #13
    Join Date
    Aug 2002
    Location
    Seattle
    Posts
    5,512
    nginx:

    - Better performance
    - Can setup wildcard reverse proxies
    - Free

    Litespeed:

    - Nice GUI

  14. #14
    Join Date
    Mar 2006
    Posts
    74
    And yes php is configured ok, its suphp and server is on cloudlinux.
    you should use php-fpm , it's better

  15. #15
    Join Date
    May 2009
    Posts
    426
    Quote Originally Posted by nwmcsween View Post
    Sadly it wasn't fully finished.

    I can benchmark Lighty again if someone provides a test server. All KnownSRV's servers are in production and I wouldn't want to benchmark on them.
    KnownSRV.com - Fast, Quality and Secure MANAGED VPS with fast SSD drives in h/w RAID10
    Full network and hardware customization possible - let's talk your needs. Optional 500gbps DDoS protection!
    █ PayPal, Credit Cards, Skrill, Payza, PerfectMoney, WebMoney and BitCoins accepted

  16. #16
    Quote Originally Posted by SceneSRV View Post
    Sadly it wasn't fully finished.

    I can benchmark Lighty again if someone provides a test server. All KnownSRV's servers are in production and I wouldn't want to benchmark on them.
    I know for a fact that lighttpd 2 beats apache, nginx, cherokee and maybe even litespeed (haven't tested it), only problem is lighttpd 2 is still in devel

  17. #17
    Join Date
    Apr 2000
    Location
    Brisbane, Australia
    Posts
    1,557
    Quote Originally Posted by jon-f View Post
    If anyone has a litespeed server that either throttling limits are disabled or you can whitelist the ab test server get with me to do some benchmarks. In my first bench i simply tested whmcs status page on a host I knew was running litespeed and compared to the company I worked for. So I think his litespeed throttled connections or something. But here is initial results.
    so benchmarked nginx vs litespeed from 2 different servers with different hardware configuration ? doesn't seem it would make a valid comparison ?? Remember if the site you're running apachebench against also utilizes mysql end, how mysql is configured and tuned between the 2 servers will also factor into performance.

    Unfortunately, no experience with cloudlinux OS as I am solely CentOS fan so can't comment.

    I've only done local server tests in non-whm environment (no spare whm server to test with yet) http://www.webhostingtalk.com/showth...70#post7267670 and http://www.vbulletin.com/forum/entry...itespeed-Cache. My main purpose was for vbulletin and wordpress performance as i use both professionally and for personal sites.
    : CentminMod.com Nginx Installer (Nginx 1.11, PHP-FPM 5.4-5.6, 7.0 & 7.1, MariaDB 10 + ngx_pagespeed + HTTP2 + lua) for CentOS 6 & 7
    : Centmin Mod .09 beta - Nginx HTTP/2+ Letsencrypt Free SSL (beta testing) support

Similar Threads

  1. Nginx Admin: The Free cPanel Nginx automated installer Plugin
    By JohnCS in forum Software & Scripts Offers
    Replies: 13
    Last Post: 05-21-2011, 11:45 AM
  2. Nginx Or LITESPEED on directadmin
    By mixmox in forum Hosting Security and Technology
    Replies: 2
    Last Post: 01-18-2011, 09:11 AM
  3. Is Litespeed any better than Nginx?
    By continuation in forum Dedicated Server
    Replies: 22
    Last Post: 05-30-2010, 05:38 PM
  4. Hiring: Litespeed to Nginx convert services
    By x264-BB in forum Employment / Job Offers
    Replies: 1
    Last Post: 04-22-2010, 11:45 PM
  5. Your Computers Benchmarks (PC and 3D)
    By BrianF in forum Web Hosting Lounge
    Replies: 9
    Last Post: 07-30-2002, 10:33 PM

Posting Permissions

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