Caldera's Almost-Linux Skips The Linux Kernel

Cassivs points to this UnixReview article, which says "Caldera has released Open UNIX 8, which includes a complete GNU/Linux distribution, except that it runs on the SVR5 kernel, acquired from SCO. It uses the same packages as Caldera's OpenLinux 3.1. It should scale much better, and provides a commercial UNIX kernel with the ability to natively develop GNU/Linux applications." It sounds like a non-Linux kernel has advantages on certain hardware, even running exactly the same software otherwise -- I wonder how long that will be true. Caldera has talked about this product, with it's Linux Kernel Personality, for a long time, and this is an informative review for anyone following it.

Comments

GNU UNIX

[GigsVT]

What is the RMS compatible way to name this?

GNU/UNIX?

I guess GNU really IS UNIX after all!

GNU....

[lga]

GNU's Now Unix?

-- Steve

Linux MODE

[alexhmit01]

As clearly stated in the article, there is a linux mode. After you type in "linux" is does things from the /linux partition and runs in Linux mode, including binary emulation. It can run and develop full Linux applications within the session.

It sounds quite impressive.

Will openunix support most x86 hardware?

[Billly Gates]

The linux kernel supports large amounts of pc, alpha, and sun devices with a huge amounts of drivers. This one of the reasons why linux is hot and openserver never really took off outside the bussiness world. All the other unix vendors run on expensive proprietary servers with only a limited amount of devices so the kernel can support them. This kernel may or may not be better then linux but I am sure you will need to run it on approved hardware for great device support.

This is the problem running alternative os's on x86 platforms.

Caldera needs to hook up with particular OEM's who want to create mission critical using openunix. If caldrea can't convince OEM's that openunix will sell, then the OS is doomed. No sane IT manager would buy an OS without approved hardware. Most Linux servers for example run on Dell or Compaq systems that are linux approved. This issue will not go away unless the kernel is opensourced and the gnu community can write drivers for it. I also hope it doesn't use the openserver libraries. Microsoft made Xenix as proprietary as possible and openserver was based on Xenix. Compiling gnu apps might be difficult. Hopefully OpenUnix's libc libraries are those from AT&T.

Anyway Caldera is dying. Its a shame because it was my first linux distro. Caldera was ahead in the linux and dos game but they did not have good marketing. Redhat totally took over. Also Ransome love's comments on the evils of gnu and opensource did not help. Client access licenses also hurt it severly. Caldera just got rid of one of the main arguements of using linux instead NT or Unix. The cost and licensing issues for small projects on limited budgets is what linux's key market is. Also linux runs on cheaper hardware. If Caldera keeps this up, then the arguement to use OpenLinux or even OpenUnix is moot. I chose Redhat for my servers thank you.

Re:Will openunix support most x86 hardware?

[warpeightbot]

No sane IT manager would buy an OS without approved hardware. Most Linux servers for example run on Dell or Compaq systems that are linux approved.

Doesn't have to be.

Any IT lead who's been paying attention knows that you can put together any old white box solution and as long as you use stuff off the hardware HOWTO it's pretty much gonna work... Oh, sure, if you're doing this on a massive scale, getting a batch of ProLiants or PowerEdges is the best way to get hardware support (and not have hardware support tell you to upgrade to the latest version of You-Know-What...).. but for small to medium sized stuff, go see your buyer with your laundry list, wait two days for FedEx, grab the stuff from Receiving, spend a couple of relaxing hours getting your paws in the hardware (you did remember your anti-static wristband, right?) and away from the CRT, and poof, time to load your kickstart CD. Twiddle the BIOS to boot off the CD, F10, go get coffee, kabam. New Linux server for cheap cheap cheap.

I mean, you do HAVE a couple hours you can take off from reading Slashdot, right? your automation scripts are up to date and will beep your cellphone if there are problems, right?

--
Something the PHB's have never figured out is that a good sysadm is first and foremost a lazy-ass s.o.b.

I have to wonder...

[trilucid]

With regard to the kernel itself, what are the ramifications of this structure security-wise? While I know this particular kernel has been around awhile (hell, SCO is ancient), it makes me curious.

To me, one of the primary advantages of using a Linux kernel is the "many eyes" approach to security. While I appreciate the fact that the distribution using a full suite of GNU/Linux utilities and such, I'd be somewhat apprehensive about the kernel itself (stability through age aside).

Anyone who has any insight into this, please reply!

the grass is always greener on the other side

[mj6798]

If you want reliability, use the user space NFS server--it has been around for years, and it can be restarted easily. Or, you could, of course, just keep running a 2.2 kernel until whatever bugs you in 2.4 has stabilized.

As for Solaris, its record is hardly stellar. For example, Solaris NFS for many years had a bug that would randomly replace blocks of data with blocks of nulls in big files (people often spent weeks trying to figure out what was wrong with their software until they finally traced it to Solaris). There have been memory leaks driver problems, and backwards incompatibilities with Solaris. Most production users of Solaris are a couple of years behind the releases in order to avoid the bugs in the new releases. And many people never wanted to switch from SunOS to Solaris at all (I think we are still running some SunOS machines).

As they say, "the grass is always greener". I can tell you from many years living with SunOS/Solaris that Linux isn't bad in comparison.

Not much new

[alsta]

I maintain two SCO UnixWare 7.1.1 servers and they do run like dogs. Linux is much faster and arguably better. Granted, UnixWare has VxFS included, but these days when we have JFS and XFS ported to Linux, that's really a non-issue too.

The SVR5 implementation of UnixWare had to be rehashed in order to increment version numbers and issue a new product. So Caldera slaps in a few things:

1) Fix apparent bugs
2) Attach more integrated GNU packages to the main distro
3) Update the Linux emulator to support more system calls

But this does not in my opinion warrant a whole new major version number. They have done pretty much nothing else with the distro. I will not be upgrading the 7.1.1 servers here as they will be phased out, but also because there is really no reason to.

Caldera is doing nothing but reselling the SCO product line. Of course they are, since there is an installed base and they can charge the same as SCO charged for it. The Linux business is, just as somebody very insightfully said, hurting them. They can't demand license fees, because all one has to do is to go to RedHat. They can't fix their distribution too much, because they become incompatible with RedHat that way. They can't charge much more than RedHat for their distro either, because not that many people would buy it. Perhaps OpenLinux is a bit ahead of its time? Probably so, but that doesn't solve their problems.

I think Caldera is realising that it has at least a temporary cash cow with SCO and thus tries to get its moneys worth. It seems to be forgetting that the business practices of SCO practically brought them to their knees and Caldera is just walking in their foot steps. The only difference is the name. Which I by the way find is pretty lame. There is nothing Open about SCO UNIX and there never was. The development kit is aged and not very good at all. SCO managed to sign some contracts with third party vendors to include some apps with their distro. These are still coming for Open UNIX. Compaq has an agreement with SCO (Caldera too?) as an OEM. This means that a customer can buy a Compaq server with Open UNIX rather than Windows NT/2000. But the sales of these are very slim.

If I was on the board of Caldera, I would swiftly make some changes in the licensing schema of Open UNIX. Granted, it can't be GPLed because of thirdparty proprietary code. But I would definately make it more available. They charge for a media kit ($65) and only give a single user license to non-commercial use. That is an enforced license which means that the system only accepts one concurrent login. This is useless for the hobbyist, so they aren't spreading the word. They have to make UNIX as exciting as Linux in order to prevail. Sure, charge the big corporations license fees. But the small businesses and home users who want a UNIX server should have to pay nothing for it.

Caldera can't do this because they would lose money that way. Some old SCO shops aren't that big and would then fall under the clause of a free OS. That means that the distribution may go up but it doesn't give them a larger revenue stream right away. Look at Sun and the free Solaris offer. Lots more people use it now, but Sun can't start charging for it unless it's for a huge server (8+ CPUs). Sun has revenue from other things, mainly their hardware. Caldera only sells software. A big catch 22.

If Caldera continues like this, they will either have to sell of the SCO division or perish. What if they can't sell it off? Will SCO UNIX become abandonware or can it be opened before they close the doors? Will SCO UNIX (read SVR5) die and be buried because it can't be opened? Would be a terrible loss.

Then there is the conspiracy deal. What if Microsoft would buy the UNIX copyright and codebase? They could stash the code in the trash can and be done with that threat. Sun could have its license revoked or perhaps be charged so much for license fees that they couldn't maintain SVR4 Solaris anymore. I know they aren't paying any royalties now, but surely that could be circumvented... That would be the true extinction of UNIX as we know it. Good in one way, but bad in others.

What if Sun bought the copyright? Or IBM? Whoever will own it in 10 years, will the codebase be opened? Who knows, but I will be following this subject over the next few years.

Alex

I've tested OpenUnix 8.0

[mikethegeek]

In my job as a QA tester, we run many different NOS's for compatibility. OpenUnix 8.0 is virtually no different from Unixware 7.1.1, other than the fact that Caldera logos have replaced SCO ones. Big freaking deal.

The OS is still cumbersome to install, and far less user friendly than any Linux distro. Of course, SCO is a true enterprise server Unix, which is it's sole (vanishing) advantage over Linux.

I was extremely disappointed that 8.0 lacked ANY improvement in user friendliness, which is the major thing I expected Caldera to bring to SCO. Caldera's Linux distros (which I also test) probably have the best installer of any Linux (though I really like TurboLinux's text mode installer). Also, the SCO shell lacks all of the user-friendliness the GNU BASH shell has, which makes it one of the hardest command lines to master (particularly when you first Unix exposure was the friendly BASH command line).

OpenUnix 8.0 is nothing more than a "slap our new name on it and get it out" to generate new revenue release. I'd advise anyone running 7.1.1 to wait until 8.1. Unfortunately, licensing isn't the ONLY one of MS's business practices that Caldera is imitating.