Slashdot Mirror
Novell to Make Linux Robust and Reliable
An anonymous coward writes: "It seems the folks over at Novell have the answer to making the "immature" Linux OS more "robust, reliable and scaleable" according to this Computer Weekly article. We have a lot more problems to use and keep running our NetWare 5 and 6 servers at our University than we've ever had with any of our Linux servers. I can't wait for Novell to help us out here."
> Ultimately they just got crushed under the Microsoft marketing machine.
Definitely. The fact that Microsoft added certain features that people want to their server OS has nothing to do with it. And we all know the only reason people use Apache is because it's free. </sarcasm>
Yes, MS is far from perfect, but I don't think Novell is/was perfect either. You make it sound like Microsoft is the only reason nobody cares about implementing Novell anymore. Nobody implements Xenix either...
PC, Nice word-wrapping. My browser wasn't doing a good enough job, so thank goodness you did it for me.
And he hasn't done SHIT for Linux. Check out what he did to his last company Cambridge Technology Partners.
If you think that SCO dude was a piece of work, this guy is even more of a cheezy fuck. What Messman really means is that he's going to put his marketers on the job of fixing up Linux right away, pump up the Novel stock a couple dollars, bail out, then watch the company spiral into the mud from his ample parachute.
We all remember when lan filesharing meant Novell. Their protocols were fast; their server was solid. They used to brag about the number of assembly language instructions between the time when a file request hit the server and the time it was being sent out.
Part of this speed came from having a very simple, unprotected operating system. Any process on the server could bring down the entire server. Novell's code was very well debugged and very stable.
And then networking started to mean something other than just filesharing. People started developing client/server applications to run on servers.
The company I worked for developed
NLMs (netware loadable modules) back in the day. It was a pain in the ass. Our code had to be flawless, because a single mistake would "abend" the server, taking down not only our services, but the lan filesharing, and everyone else's services too.
Mind you, we tested our code, we did everything we could to make it flawless, but that's a difficult standard to obtain in a complex piece of code.
Novell eventually tacked on some memory protection to the OS, allowing some NLMs to run at ring 1. But it seemed like too little too late.
Developers were realizing that it was a lot easier to develop and deploy server code on protected operating systems (Unix and Windows), and the speed bonus that Novell got by writing a down and dirty operating system was becoming less critical as machines got faster.
The same thing will happen with Palm OS vs Windows CE and Linux for the handhelds. The miserly memory handling and power consumption features of Palm OS will not be needed in future devices, and modern operating system features will win out.
# (/.);;
- : float -> float -> float =
All I can add is the following :
I have used Windows, Linux and Novell on the File server side.
Give me Linux over windows for up-time, and Novell over both for up-time, scalability and
ease of admin.
The early version of Netware 4.x blew chunks,
but from 5.x on I have not had any problems.
I have 3 netware servers currently up and running
for over a year without any problems whatsoever.
This includes file, print and groupwise email.
I am also running both IP and SPX.
When their edirectory product integrates better on the Linux side
I would be tempted to move to everything to Linux, but until then,
screw the marketing dweebs who blather about market share, etc.
Better to use what works than what is popular IMO
As a bonus for not running the "hot thing" in NOS,
our servers have never been affected by virus problems.
KDE is quite equal to Windows/Macos. The installers are up to par as well.
What Linux's core market is is in the server room. WIthout a directory service it can not totally replace Windows. With Windows in the picture its more tempting for IT managers to standardize on Microsoft.
Novell is great to administer and does things that Unix/linux can not do.
Its not an apps server but really a network service provider.
Linux is missing directory services, enterprise journaling( not the ext3 hack), administrative utilities, and the ability to manage profiles for desktops easily.
I think this is what Novell might of meant. There is more then just not crashing.
Novells problem is they use an ancient kernel that is not memory managed that well. This is from the 286 days and it does crash. Reliablity is a problem. WIth a linux kernel that can be fixed.
The networking services addition from Novell would rock and they would cream Redhat or any other vendor if they had it. Deffinetly Novell Linux would be the pick from any corporate manager for this reason over redhat.
http://saveie6.com/
Oddly, I'm currently at BrainShare and he did not make those sort of comments during his keynote this past Monday. In fact, he even made fun of Scott McNealy's penguin suit and set a positive tone about Novell's interaction with Open Source. He also made a point about Novell being slow to listen to market changes and how that was being changed (he used IPX as the example).
The actual product roadmap came from Chris Stone, the vice chairman. Unlike the arrogant comments by Messman in the linked article, Stone seemed much more humble. He talked about the various Open Source technologies shipping with the next version of NetWare (6.5), including MySQL, Tomcat 4, Apache 2, and PHP 4. Finally, he announced that Netware 7 would run either the Netware kernel or the Linux kernel. He made it clear, however, that Linux was the ultimate destination. There wasn't any dismissal of Linux, especially since they expect to base all of their products on top of it.
Does Novell have anything to contribute?
Well, they claim that they've contributed back many improvements to PHP, Apache, and MySQL. Some, they said, were still forthcoming but that they would be available to the larger community.
As far as their products go, they still make a surprisingly large number of good ones. Many of the services that do run on Netware, including iPrint, iFolder, NetStorage, etc. would be a welcome addition to any operating system. eDirectory's already available cross platform, so nothing is new is gained there. Provisioning and user account management with Netware/eDirectory is still superior to many alternatives and makes administering a large number of users very easy (especially for support folks).
So, I think Jack Messman's comments are regrettable, but I don't really care what he thinks. I'm here at BrainShare to speak with the developers of the products we use and they, almost universially, get it. In nearly every session I've attended, they've highlighted solutions available from Freshmeat, SourceForge, CPAN, and others. I think it's especially helpful since most of the attendees here are not Slashdot readers. They're old school Novell admins working in a range of industries, from very large corporations to small business consultants. Despite the bravado from some CEO, Novell's participation should be welcomed and encouraged. After all, if they're contributing something useful, why not?
I worked for Novell until about a year ago, and I have to agree with a previous poster who said that this strategy was all about customer retention. Show customers a direction towards Linux, a little bit of open source, and toss in some buzzwords and customers might keep their license agreement. It's a good strategy financially and not unlike what Microsoft has done in that arena.
Netware's list price is over $100 a seat. Even if MySQL, Apache, and anything else ported over worked perfectly, no one is going to buy a linux-based Netware as a linux replacement. eDirectory runs about $2 a seat list and has been running on Linux for a long time. The announcement of a free UDDI server is nice, but I don't see long term how Novell will get a piece of anything in the Web Services space with that. It's more of a developer tool, and Novell isn't a developer tools company, they make money selling to big corporations. Yes, they recently acquired an app-server company, but that's an ever worse competitive mess than the LAN arena.
I think Novell's main problem is too many products. There are still just as many products at Novell as there were two years ago, but there are probably half the engineering staff to maintain them. Products like iChain and DirXML are incomprehensible to most people, and too narrow in scope and low in sales when most of their competition are rolling their products up into big do-all authentication suites. Also, there haven't been installation or adminstration console standards at Novell for years and years, so getting two different products from different groups running is quite a challenge.
While I'm a little bitter over some of the specifics of my departure, I think overall Novell has good people and still has a large user base. It's hard to turn a big boat like Novell towards new technology when the old stuff is still raking in hundreds of millions of dollars. Hopefully this won't end up like the two previous major efforts towards *nix, the first being the purchase of AT&T Unix and the "SuperNOS" strategy and the second being a major alliance with Red Hat that never really went anywhere.
Good luck, guys!
- "When you want something with all your heart, the entire universe conspires to give it to you" -Paulo Coelho
Remember, you can't make everyone happy all of the time, so why bother trying? I wish linux distros would grow up and find their market segments instead of trying for everything. Even Apple has recognized the need for separate desktop and server "distros".
Don't be stupid. Server operating systems and desktop operating systems are similar. The fundamentals are the same. The differences are in the fine tuning.
Debian produce probably the most stable (in all senses) Linux distribution. They don't have a Server distro and a Desktop distro. When you install the operating system, you choose the components that form the type of system you are creating. What is wrong with this?
Returning to your final comment about Apple, their strategy of releasing two different OS's for desktop and server is a marketing decision, not a technical one. There is no difference between the two OSs so huge that the decision cannot be made at install time, instead of being brought forward to CD-burning time.
Yours Sincerely, Michael.
I dunno abt you, but when I want to install something, all I need to do is "apt-get install foo"
This downloads the software, all its dependencies etc., installs it and is all done in seconds (without subjecting me to any "wizards")
I'm not wild about RPM, but I certainly think apt is far superior to anything Windows has.
RedHat's "printtool" makes it easy to install printers too. (Its GPLed of course, so its available on most distros)
When it comes to getting pictures from my digicam, and samba, however, I agree with you. Ditto for hardware support, although that's really a vendor problem. So long as you have third parties writing drivers, its always going to lag. For example imagine how unusable windows would be if you were restricted to drivers that come on the WIndows CD. Get hardware that's newer than the windoze version, and you're screwed.
I remember back to when Novell bought UNIX.
The pitch was:
UNIX is filled with cryptic things like grep. We'll be getting rid of grep and all thoes cryptic things.
So, personally, I'll just sit and wait for those earth shattering improvements in whatever Linux fork they are going to do.
Hmm...funny, but I've checked the mlist.linux.kernel list, and I don't see any Novell staffers contributing to the kernel.
How are they going to mature Linux? Make lots of outlandish marketing promises to the general public? Oh wait, no, it's Novell we're talking about here.
Ruby on Rails Screencast
Seems to be their claim to fame these days.
They are simply, unnecessary, and have been for a good few years now. It'll take them a while to work through their financial reserves, but eventually they will wither and die like the dinosaurs they are.
Open source is like the asteroid which smashed into the earth destroying the ecosystem the dinosaurs needed to survive. It's literally pulling the financial flora out from underneath them. What will rise up in the aftermath? Mammals. Small, fast and flexible companies which can thrive on resources which wouldn't have fed a dinosaur for a day.
Just keep out of their way as they go through their death throes.
Government of the people, by corporate executives, for corporate profits.
Novell was slow to support IP because they didn't need it. IPX was working just fine, thanks. They did definitely miss the boat with web-integration and bordermanager is a steaming pile of shit, but for file and print services, NDS absolutely rocks. And for application delivery, zenworks is untouchable.