Talk:G-WAN vs Nginx
Quite honestly - while wikivs might not be a really serious site, this g-wan advertisment has to be a bad joke. As the author/advertiser of the initial "article" (who is more than likely the g-wan-author himself) does not hesitate to revert or relativize valid points on nginx' side of the VS (which some relatively neutral author added later on, as there was no pro-argument for nginx at all in the initial revision), everybody should really question this article and every single claim. Also, while citations to nginx' shortcomings point to 3rd party sites, citations to g-wan's "advantages" only point to the g-wan site.
"Quite Honnestly" wikipedia does not look like more serious than wikivs - especially since you consider the systematic censorship on wikipedia. And, if you refer to the "Lighttpd_vs_nginx" comparison on wikivs, it seems that the nginx promoters do not fear the practice of one-way comparisons. The difference, this time, is that Lighttpd is not the victim. Now, if you are ready to question every single claim you are more than welcome - the G-WAN author does not seem to fear COMPARATIVE benchmarks (nor publishing the benchmark source code to help you prove him wrong) so the claims look more solid than those made out of the blue - whoever made them on so-called 'independent 3-party sites'.
G-WAN isn't being censored on Wikipedia, it's being removed because it isn't noteworthy and frankly a worthless piece of trash. Wikipedia isn't an advertising platform and WikiVS shouldn't be either. This whole article needs removing. A piece of shit like G-WAN is never going to have any users so it's utterly pointless even discussing it.
The G-WAN author clearly feels the need to brag about his "superior" design to anyone foolish enough to listen. Benchmarks do not tell the whole story and can be manipulated to support whichever claims you like. http://kristianlyng.wordpress.com/2011/03/16/the-many-pitfalls-of-benchmarking/
It seems that this page is like a joke, PHP uses 80mb of ram? Do you have started it with all pecl modules loaded in? My embedded own-compiled system with nginx+php+dnsmasq+cron+a-lot-more used 17mb of ram! More, the author isn't considering that an application needs to be written by someone using some tools ... i can found a lot of stuff for php, java andh c#, i can find a lot of editors ... i can use visual studio + vs.php too ... gwan/c scripts misses that stuff. More, write a complex and large web applications (author claims gwan to be an application server) in C requires a LOT of time, while using php, java or .net reduce a lot development and debugging time and let you to reuse existing stuff. Before write an article like this, just use your mind!
i'm the one added the note "this is junk": sorry for the modification to WikiVs page related to GWAN/Nginx comparison but this article is full of technical inaccuracies. If this article is a comparison between two software the paragraph "This article" is a non sense, the g-wan author should talk with wikipedia admins and not just put this stuff, that isn't technical at all, there.
More, the initial paragraph of Design seems a slogan because just say "gwan has this, this and this too while nginx has nothing special and is complex to configure" but, really, this is a non sense because gwan is an application server, like the author states, and nginx is a web server and this means that while gwan doesn't need to host misc stuff (just static files and c scripts) nginx needs to handle everything so it need configuration. So, i repeat, this is a non sense because these software can't be compared in general because aim to do different things. Related to scripts paragraph, while it's true that C scripts outperform any other thing, nginx doesn't support perl to let the user to write entire web application but just to handle dynamic configuration. Continuing, i don't know java requirements but PHP needs around ten mb of memory with standard configuration and a couple of extensions and C#, on mono, needs something more than php (just because needs to start xsp/mod-mono-fastcgi).
Related to libraries, the same ... the article author miss that nginx needs memcache to cache responses and a module exists, while gwan can link to memcache just to let application developer to use it ... it's the same? No. Can be compared? No.
Related to security, but i can say only that the gwan author is a "dumb" ... gwan is more secure because it's less used? Really not, at least it's less tested than others but not more secure. If the G-WAN author, windows, from the young age, should have been really secure but it's not ^^.
I just want to skip the "performances" paragraph because a part of one benchmark the other stuff is just. Related to GWAN/IIS comparison, i want to say that gwan miss a lot of stuff, as stated on the download page (ssl, tsl, service mode, virtual host, http compression, auth and a lot more stuff) so really shouldn't be compared to anything.
Really, i don't want to continue ... but as i've wrote ... this is junk ... it's not a comparison.
It is enlightening to see how many erroneous facts you feel necessary to publish to downgrade G-WAN (which supports a daemon mode, virtual hosts, HTTP compression, etc.).
Independent comparative benchmarks have been done by Nicolas Bonvin (Google: "Nicolas Bonvin + G-WAN") with half a dozen of other Web servers - and G-WAN was much faster than all on Nicolas' modest i3 laptop!
And, in addition to ANSI C scripts, G-WAN also supports Java, Objective-C, D, Lua, Javascript, Google Go and Python.
Further, Application servers (GlassFish, Tomcat, etc.) are always slower than mere Web servers because of the additional interfaces needed to generate dynamic contents.
Therefore, having an Application server (G-WAN), beat by several orders of magnitude mere Web servers (like Nginx) is not a small feat.
