User:Boborok/Archive/Talk:Zeus Web Server

Why is there even a debate about UNIX and UNIX-like? From wikipedia's own article on Unix-like:

'A "Unix-like" operating system is one that behaves in a manner similar to a Unix system, while not necessarily conforming to or being certified to any version of the Single UNIX Specification. '

Examples:


 * 1) GNU – Other systems based on GNU are often identified by the kernel they use:

* Linux (or GNU/Linux) Is the user who keeps defacing this article an employee of SCO? GNU/Linux is *not* UNIX. Rather, it is an OS that looks and 'feels' like a UNIX OS, indeed 'behaving in a manner similar to a UNIX system'.

If Boborok persists in defacing this article, then a sysop must be made aware to control their abuse. --Toph3r 19:45, 30 April 2006 (UTC)


 * Sorry, GNU is not even an operating system. For example, a very important aspect of UNIX-ness are the supported syscalls. Syscalls are provided by the kernel. The Linux kernel is not GNU-based (it contains exactly 0 lines of GNU code). I am not defacing the article, but trying to clean it of nonsense. Read this http://www.unix.org/what_is_unix.html (Boborok 21:41, 30 April 2006 (UTC))

-- ARGH!! GNU *is* _part_ of the Operating system! It uses a variety of kernels (whether it be open-solaris, Linux or Hurd) to perform the functions of an OS. Stop quoting unix.org btw - it doesn't even back up *any* if your assertions!


 * What is more authoritative on UNIX than unix.org then? It's run by the Open Group who own the UNIX trademark. And yes it states right there that a system doesn't need to contain any AT&T code to be considered UNIX. (Boborok 02:41, 1 May 2006 (UTC))

Incidentally, I never stated that GNU was an OS in the first place. GNU/Linux is the correct term to use when referring to a Linux kernel based OS, utilising the GNU user-space toolset. I suggest you read the wikipedia article on GNU and GNU/Linux Naming Convention. Time and time again, you are unable to use (or indeed, *know*) the correct terminology. Read this:

http://groups.google.com/group/net.unix-wizards/msg/4dadd63a976019d7?output=gplain

Stallman himself states that GNU will be 'Unix-compatible' (now referred to as UNIX-like), and will utilise its own Kernel (which was hurd, but that failed) - however, Linus came along and wrote Linux, which the GNU project plummed in as their OS kernel.


 * This is Stallman's contorted point of view, and Wikipedia is supposed to be neutral. (Boborok 02:41, 1 May 2006 (UTC))

You are wrong when you state that Apache performance isn't compared -- you stating this shows your lack of knowledge. Apache isn't used for SPECweb submission tests because it performs *woefully*. ZWS is used extensively for SPECwebbing because it, consistently, performs fastest on (usually highly scalable) platforms. The Sun Web One test, if you look, is conducted on hardware FAR out of the range of comparable ZWS test hardware.


 * Why is Apache used on about 70% of all the world's websites if it performs "woefully"? Again, this is just your biased point of view. (Boborok 02:41, 1 May 2006 (UTC))

And, now you've omitted 'Unix' from the list of supported platforms - argh! You've specificed, BSD and Solaris - but what about Tru64, AIX, HP-UX, IRIX - those are all supported UNIX platforms ZWS runs on. See, if you had just kept 'Unix and Unix-like' in the article, there wouldn't be this ambiquity. It's time to get some official moderation on this, as your edits are lacking any amount of clue :( .--Toph3r 00:30, 1 May 2006 (UTC)

Some points:

1. Apache is acknowledged in the industry as having pretty sub-standard performance without doing a *lot* of tuning to the product. Even then, it's model of forking, and now pre-forking, just doesn't cut it most of the time. This is why load balancers are routinely deployed against Apache. It's not tested under SPECweb because the companies testing *know* Apache performs so badly. Thus, this is why Zeus Webserver is routinly used.


 * And this is why SPECweb benchmarks are useless. (Boborok 16:54, 1 May 2006 (UTC))


 * Sorry, *why* exactly? Benchmarkers are free to use whatever webserver they wish with SPECweb -- the fact that they choose to use Zeus should tell you everything you need to know.  Apache clearly doesn't cut it regarding high-performance web-site serving.  It's a great product, but their forking model has held it back for years.


 * (above unsigned comment by "User:Toph3r")


 * Apache also has a "worker" MPM using threads for those who don't wish to use the preforking model. But there is a very good reason why Apache is using the model it's using. When serving big files to many concurrent slow connections (as in a real workload), the kernel is free to reorder I/O, whereas with a single-process event-notification based model (as in Zeus), this workload can lead to the hard disk seek time becoming the bottleneck. This could be solved by using AIO, but is Zeus doing that? That's one reason it would be important to compare against Apache. (Boborok 20:18, 1 May 2006 (UTC))

2. Apache runs 70% of the worlds websites because it's *free*.

3. UNIX.org states that Linux is *not* UNIX. I cannot believe you would ever think to make this assertion!

4. How can Stallman's view be 'contorted' when partly wrote a lot of the OS I, and many others, use on a daily basis? If he says his OS is 'Unix-compatible', then it *is*. The very reason GNU was started was as a FREE alternative to existing, propietary, UNIX systems. Thus, GNU/Linux is *not* UNIX. Indeed, what does GNU stand for? GNU's NOT UNIX!

5. You are perhaps the most clueless person i've come across who regards himself as an authority on computer science.

6. The article's now been protected and will stop your vandalism -- and that *is* what you've done, vandalise this article.--Toph3r 07:57, 1 May 2006 (UTC)


 * I put the protection on it, but apparently it doesn't work. (Boborok 16:54, 1 May 2006 (UTC))


 * You put the protection on it, to stop your own vandalism..? Look at the article before you changed it - now look at it. All you have accomplished is highlight your severe short-comings regarding Unix/Linux debate and added, information regarding publishing of benchmarks.   Will you leave the page alone now?  Can't you go and vandalise other articles - you have a history of doing so according to your talk page. --Toph3r 17:43, 1 May 2006 (UTC)