Slashdot Mirror

You make very good points. Apache should get it's shit together and become more user friendly. Fast. I find configuring Apache 2.2 a pain and I know most configuration options by heart!

I have been more than happy with LiteSpeed as a web server. Fast PHP built-in, easy web interface, and the server is lightning fast. If you're tired of Apache I suggest you give it a try. Not free as in speech though (only as in beer) which I find not problematic.

PHP and a DB persistence layer should get moving too. The standard php-mysql is just not going to cut it. It doesn't even support prepared statements. If anything is behind Rails' success, this is.

I fear that Apache+PHP have become inert and complacent with their own succes. Which they have, LAMP is still a rock solid platform to deploy apps on. But in tomorrow's world LAMP will become obsolete quickly if it doesn't evolve.

But isn't WEBrick only intended for development or very small sites ?

But isn't mongrel intended for very few connections and must have a load balancer in front of it ?

What are real life experiences.
Lighttpd with fastcgi: http://www.lighttpd.net/ ?

Lightspeed: http://www.litespeedtech.com/products/webserver/do wnload/ ?

Or from other people experiences that I've been reading about go with Litespeed http://litespeedtech.com/

IIS is also arguably faster, as it's running on a single architecture, on an OS designed by the same developer.

I don't know if these numbers are trustworthy, but at least its another datapoint.

they want to make sure everyone is nice and compliant about upgrading when they decide to take httpd over to java like all the other java kool-aid they are selling --- Maven is Jonestown, lets all program in XML because its standard! Cultures breakdown when there is too little disent and questioning of authority, the apache foundtion is headed in that direction.

Lets move on, SOA and all that, most people don't need any of this mod_* crap and could use:
thttpd he has other servers there, too and http_load.
lighttpd I'm moving to this sweet little server for most apps and the home site runs ea php and ruby on rails
AOLServer like OpenACS runs on
Boa
fnord from our boy who did the (in)famous benchmarks
Cherokee I root for this one for some reason.
gatling
cthulhu
yaws in erlang, should support more simul. connections than the unlying OS can support.
dhttpd
Litespeed check out their php benchmarks
thy
roxen
mini-httpd never tried this one
xitami I have a intranet server running for 5 yrs (without upgrading xitami) on xitami Solaris, simple, small, easy to admin, never dies max uptime was 1000 days+.
eddiefor complex load bal and geographic distribution
hiawatha

And for the love of god, please at least design your sites to get their images from images.mysite.com if possible so that you can use a non-bloatware web server to server the images, reserving horsepower on your apache server for stuff that actually _requires_ some features of apache.

http://www.hcsw.org/awhttpd/ updated on 12-06-2004
http://www.norz.org/zawhttpd.html
http://cr.yp.to/publicfile.html

• that they took a resource that was built by community input on the basis that it was free and open and then closed it off and started charging for it - the RBL (tm) and family.
• They have a "club" for the elite in the DNS world that they charge $4,200 - $50,000 to join. oarc.isc.org - Why have ISC "own" this and not have it as separate org that everyone including ISC is a true peer? Because ISC wants control, that's why or else they could easily have formed a separate non-profit that elected a board and was not under the control of one entity.
• They tried to make money by offering exclusive early access to security patches to those who paid - this was viewed as a huge breach of trust in the Open Source community.
• They run their money making companies separately so that ISC looks poor

Paul Vixie created Nominum as a separate company then hired them to create BIND 9, Nominum offers support for BIND at 50K+ for most companies that have received quotes. Nominum also sells a higher performing version of DNS server, does that mean that they held back the best stuff from the Open BIND version to use only in the closed source version? Maybe and that is their right if they so choose, but at least most developers are up front about it, like Litespeed our free stuff is good and the pay stuff is better - very up front and all the info in one place.

Here are some quotes from a marketing PDF from Paul Vixie's money making company (Nominum) about the software they created for the Open Source company:
[Note that I find it hilarious that Paul is having his money making company assail BIND on the same points that DJB has been for years and they (ISC and its supporters had, prior to Nominum's product release, tried to say that DJB was crazy and that DJB's modular server design was complex and modularity was unnecessary) yet they then turn around and design their money making high performance product more like DJB's design than like BIND, hmmm...]

• Scalability. BIND 9 scales poorly in terms of queries per second, size of zones, or number of zones. It was designed primarily to meet the needs of a mid-scale DNS system in the mid-1990s, not to address the expanded uses to which DNS is being applied.
• Reliability. The multifunction design of BIND allows it to be configured as an authoritative server, a caching server, or both. This results in an unnecessarily complex system. Since BIND is multipurpose, the relatively simple authoritative function is laden down with excessive code, which impacts reliability. [Ed: and security, too] Though also an issue in BIND 8, it is exacerbated in BIND 9 because of BIND 9's preemptive multithreaded architecture.
• Security. Since BIND 9 can be used as a multifunction server (simultaneously authoritative and caching), many administrators set up their DNS servers this way. DNS experts recommend against running DNS servers as multifunction to prevent cache poisoning, [Ed: only after DJB embarrassed them into admitting that, but they still downplay it in the BIND docs] which can lead to serious security breaches.
• Manageability. One of the major failings of BIND is that the server "goes deaf" during a reconfiguration restart.
• Performance. The DNSSEC extensions to DNS allow cryptographic authentication of DNS data. When these are enabled in BIND 9, performance degrades to a point that some administrators would rather not enable DNSSEC. In part, this performance impact is due to fluctuations in the DNSSEC drafts while BIND 9 was being written. In any event, it creates an unacceptable tradeoff between performance and security.

There also appear to be financial connections to a major ISP and a mailing list