Outside Looking In: The BSD Operating Systems

 
 
By Steven Vaughan-Nichols  |  Posted 2003-10-31
 
 
 

Outside Looking In: The BSD Operating Systems


Discuss This in the eWEEK Forum

A while back, someone asked me why the open-source BSD operating systems arent as popular as Linux. Its a good question. Technically speaking, the BSDs are often every bit as good as Linux. Indeed, when it comes to security, OpenBSD is the best of breed.

Indeed, you can argue that BSD is actually more successful than Linux on the desktop. Im referring, of course, to MacOS X, which is based on Darwin. Darwin, in turn, is built on top of Mach 3.0 operating-system services, which runs on top of the 4.4 BSD operating system.

However, developers usually see Darwin simply as the foundation for MacOS X and its Cocoa and Carbon toolkit-based application. Apple says it would like Darwin to become an operating system in its own right for both its native PowerPC and Intels architectures. In practice, though, Darwin has made little progress as an independent operating system.

So why hasnt mainstream BSD become widely popular? The answer is that there is no mainstream BSD. Like the many Unixes of the past, it shot itself in the foot long ago by dividing into several different variants: SunOS, Wind Rivers BSD/OS, FreeBSD, NetBSD and OpenBSD. With five major forks, not counting Darwin, only SunOS, which became part of Solaris, has had great commercial success.

This was also the fate of many other Unixes, such as Coherent; Interactive Unix; Dell Unix (yes, Dell once had its own Unix); and Consensys Unix. Until Linux came along, both proprietary- and free-software Unix developers built their own unique operating systems. How many? By my count, theres well over a hundred different Unix and Unix-like operating systems. None of which, need I add, worked that well with programs written for any other Unix variant. It was, and is, an approach destined to make sure that no single Unix could achieve great success.

Linux broke this misguided model, in large part I think because of Linus Torvalds combination of quiet leadership and a willingness to let anyone and everyone code for the project via mailing lists and Usenet. This more-open approach gathered up far more supporters than the much smaller BSD development communities. In Linux, anyone was welcome to lend a developing hand; in BSD, only the truly devoted were welcome. By being inclusive, rather than exclusive, Linus made sure that Linux avoided the problems that kept not the BSDs but all the other Unix operating systems from mass acceptance.

By adopting the GPL and GPL software, Linus also soon garnered support from the Gnu/Free Software Foundation (FSF) community. The Gnu folks may have never gotten much of anywhere with their own Gnu Hurd operating system plan, but their tools, especially Gnu C, gave Linux a big leg up.

At the same time, the BSD community had never gotten along well with the FSF because of free software religious wars over GPL vs. BSD licensing issues. This kept the BSD developers from taking full advantage of the FSFs excellent development tools.

Since then, Linux has continued to gather developer support. In time, Linux parlayed this into business support. This has snowballed so that today for every mention of the BSDs, youll see one hundred mentions of Linux.

Again, from a technical and licensing viewpoint, the BSDs have a lot going for them. I run FreeBSD and OpenBSD myself, and Ive said it before, and Ill say it again, OpenBSD is the most inherently secure OS around.

But what the BSDs dont have is the business and technical support, independent software vendor products and support needed for any of them to become a first rank business operating system.

It could still happen. But for now, youre only going to see the BSDs in businesses with in-house staffs that have mastered them, and need them only for the server basics such as Web and file serving, firewall, routing and so on.

This isnt just my two cents. On the next page, Brian Masinick, a software consultant with vast and broad Unix experience, shares his thoughts on where BSD is today and why it hasnt garnered Linuxs popularity. Since I couldnt say it better myself, Ill let him tell it.

Next Page: A software consultants take on BSD.

An Insiders Take



Software consultant (and Linux & Open Source Center reader) Brian Masinick writes:

    Ive been a regular Unix user since 1982, and I dabbled with it on and off even before that. In 1982, I worked on a project to assess the applicability of minicomputer and PC technology for massive use in the software-development community at GM. We needed to complement our software development process, which was being completely done on mainframe computer systems. It was expensive, for one thing, and any outages at all would completely knock about 500 developers off schedule.

    We looked at Unix with interest, both as a technology, in and of itself, and as a conduit of communications between PC systems in various groups and somewhat larger groups of systems. We anticipated PCs becoming the tool on everyones desk, but felt we needed more connectivity, particularly with the extremely limited connectivity between PCs and anything else back in 1982.

    NCR had an interesting minicomputer tower system based on Motorolas relatively new 68000 chipset. About a year later, we got our hands on a relatively new system from another vendor (now very familiar, Sun Microsystems). Both of these companies ran Unix software on Motorola based systems. Both had software that included BSD-based Unix implementations. It looked promising, but neither were well polished at that point in time. But I looked and learned with great interest.

    A few years later, in 1985, I had the opportunity to leave GM and join Digital Equipment Corp. (DEC). They, too, had been working with Unix, mostly because of their large customer base at AT&T and the recently divested Bell Operating Companies (called BOCs, later Regional Bell Operating Companies or RBOCs). These companies often insisted on using Unix, but generally insisted on using AT&T Unix System V. Digital quietly wrote drivers for these companies and sold them to the companies, which already had their own AT&T source code, but to other customers, Digital had been establishing customer interest in a university-developed version of Unix known as BSD Unix. Digital developed its own proprietary version and called it ULTRIX. Years later, Digital developed a hybrid version of Unix and called it Digital OSF/1, renamed it to Digital Unix, and then changed it again to Tru64 Unix, just around the time that Compaq acquired the company.

    Because of this background, Ive probably used BSD-based Unix systems much more than I have anything else, but I also worked with the telecommunications industry, so I also had quite a bit of contact with Unix System V Release 2, much older than the current Unix System V Release 4 (SVR4).

    Given that background, when I started to actively pursue an interest in trying out PC based Unix systems for personal use, I intended, from the beginning, to try out BSD based software, along with GNU/Linux software.

    As I started along this path, I ran into hardware compatibility issues. I found many of my issues stemmed from the fact that I had a Compaq Presario system, so I moved to a Dell Dimension 4100, a custom built (but aging) AMD 400 MHz system, and an equally aging CyberMAX 400 MHz Celeron system. Once I did that, I found that I was able to use both BSD and Linux systems. But by that time, I actually got to install and use much more desktop Linux software than BSD software.

Next Page: How BSD is lagging behind Linux.

BSD Lagging



Software consultant (and Linux & Open Source Center reader) Brian Masinick continues:

    Ive tried installing and testing out FreeBSD, NetBSD, and OpenBSD. While none of them are terribly difficult for an experienced systems administrator to install, compared to the progress thats been made in the past three years on GNU/Linux system installation programs, all three of the free BSD systems lag quite a bit in terms of simple, easy to use installation programs, particularly for consumers.

    For those who get past those things, BSD software, in any variety, is stable, extremely flexible, arguably better tested, more secure. At the same time, those things also mean that it tends to be less bleeding edge, slower to come out with new features, and more difficult to initially install.

    FreeBSD has a really interesting way to deal with this. The current branch of software is much closer to the leading edge, and can approach what Linux users are used to seeing. At the same time, it has a stable branch, which contains rock solid, well-tested software.

    The other BSD variations seem much more intent on providing stable, secure software thats also portable, but they, too, have adopted a way to get at somewhat more "current" software, then provide a very stable release branch. Ive not yet had a good chance to thoroughly assess all of this to compare how these branches compare to typical commercial Linux software.

    Want more BSD insights? Then check out a detailed Extreme Tech article here.

    On the Linux end of things, Red Hat, Mandrake, and SuSE dominate most of the easy to use commercial software, but recently, companies like Lindows.com, with its LindowsOS, Xandros, with its updated version of what used to be Corel Linux, and small desktop players, such as Lycoris Desktop/LX and ELX, have made Linux software offerings quite interesting and very useful, even to casual consumers.

    In the noncommercial space, the Debian GNU/Linux project has been most interesting to me, and over the past year, Ive developed more interest in it than anything else. From the Debian project, I can get a choice of either stable or bleeding-edge software, and I can get it in either binary or source form. As long as I have a good network link (which I do), I can easily download and customize my system to suit my needs and interests.

    I still intend to do a lot more exploring of the BSD implementations of software because I have so much background in BSD based Unix systems. But as someone whos used both Unix and Linux software, I can tell you that it is far easier and quicker to get into any number of Linux systems and come out with a really solid, usable desktop system. If I were running mainly server software, Id think itd be much more worthwhile investing the additional time and effort to secure an OpenBSD system rather than a GNU/Linux system. However, I do not feel that the security differences are large enough to worry greatly about, for those who dont have a lot of systems background. I think thats why so many individuals and companies (myself included) have opted to focus more on Linux systems for every day use.

Discuss This in the eWEEK Forum

eWEEK.com Linux & Open Source Center Editor Steven J. Vaughan-Nichols has been using and writing about Unix and Linux since the late 80s and thinks he may just have learned something about them along the way.

Rocket Fuel