Valid XHTML 1.0!


 

The following presentation is somewhat subjective, and this can hardly be avoided. Either way, I try to be as fair as possible, while focusing on the most important participants in our common 'defense war efforts' against that dangerous mob from Redmond...

Skip to: GNU/Linux   BSD Unix   IRIX   Mac OS X   the Future

SUN Microsystems: the commercial Father of UNIX

Solaris logo In the early 80's UNIX was already in use by a relatively small community of developers and administrators, mainly in scientific institutions, not unlike the situation regarding GNU/Linux about 10 years later (see below), but it had no commercial support whatever. In this very moment the founders of this fine company grasped an huge opportunity: to become the first commercial UNIX system vendor and so the ascend of Sun and UNIX became closely intertwined. Both 'participants' gained large advantages from this "marriage": the open standards, on which UNIX is primarily build, were enriched and extended by other freely usable and open standards invented by Sun like NFS, and Sun put the first open, truely interoperable OS on the market, opposed to the than prevailing proprietary OSs, which were virtually unable of data exchange with each other (at that time it was called SunOS; after adapting SVR4 [System Five Release 4] standards in the early 90's it was renamed as Solaris). You can call Sun The UNIX Company without any exaggeration, I think, because they did never anything else than UNIX on the OS side (recently they buyed a company, which offers rack mounted ia-32 servers running GNU/Linux...). Recognising the huge success of SunOS workstations, all other major workstation and server vendors resorted to create their own UNIX systems... and virtually all proprietary OSs in that market area died --- and this happened rather quickly in a few years. If you're interested in more details, read Suns own account of that story. Later on Sun revealed even the majority of the source code of Solaris, ownly sticking to the decision rights on it. The perspective is still good, maybe Sun will remain the only or one of very few (up to at most three, I guess) own UNIX system vendor(s) in future.

Linux and GNU/GPL: Penguins conquer the IT World

LINUX penguin In an university world dominated by UNIX systems of different flavors, with SunOS being the most widespread, a young finnish student wanted to have it too on own hardware he could pay for: he grapped a low cost Intel based PC and began to write his own UNIX kernel for this than not really supported platform (only SCO, today Caldera, had some --- expensive --- UNIX available on that ia-32 platform at the time). Some years before the GNU (GNU means Gnu is Not Unix, much irony within!) organisation began to deliver own contributions regarding improved command line tools for UNIX and more as Open Source under the famous GPL (GNU Public License). These two developments fused a little later together in todays GNU/Linux system, which complies completely to the GPL (despite Debian is the only distributor to offer this "pure" GNU/Linux and under it's correct name). And this is the major reason for its success (compare below: FreeBSD!): the license, which requires open sources, but maintains the contributors mindshare and invites others to join the effort on a par inter pares ("all have equal rights") base while prohibiting one-sided abuse (see also below, FreeBSD!), encourages all people to take part and to gain not only reputation, but also money (spared and earned) by indirect effects. That's the reason behind IBMs huge efforts and investments into GNU/Linux: after they lost in the UNIX market badly with their own UNIX (despite since AIX 4.3 UNIX 98 compliance like Solaris 7 and later, it is in many regards more proprietary polluted, than any other UNIX flavor) against Sun, they try now to overtake Sun by being more open and standard compliant than their rivals! However, for GNU/Linux is the multi billion US $ support of IBM extremely valuable; most other UNIX system vendors are contributing too to GNU/Linux by protecting and supporting distributors as well as the system itself: HP, the former DEC (today Compaq) and SGI to name the most important; other companies like Oracle and SAP are supporting Linux too. In five to ten years GNU/Linux will probably be either the dominating UNIX flavor of all, or at least one of the at most three widespread UNIX systems, which will then be the only used OSs at all.

(Free)BSD Unix --- the Daemons of Freedom

FreeBSD daemon In Berkeley some people worked on an own UNIX derivative, and with the time coming the line of BSD UNIX systems emerged from that. Today mainly three free of these are important: NetBSD with special features for networking, OpenBSD, a security and reliability enhanced (by double-checking the code) system, and probably the most widespread of these, FreeBSD, sometimes considered as alternative to GNU/Linux, but the license model is different and in the view of most observers the primary reason for the different success: the completely free license of FreeBSD enables third parties to grap any code of it without any duties of importance. And of course there are unscrupulous companies (you know the mainly guilty one), which abuse this shamelessly. Despite you could make some positive aspects out of this, for example correctly working TCP/IP stack implementations in such proprietary systems by incorporating FreeBSD code, this hinders cooperation instead of promoting it, which the GPL has done and continues to do so very good. Either way, sometimes such a license maybe of better use, than that worst case mentioned...

SGIs IRIX: graphical High End Unix

IRIX When you have seen "Jurassic Park" or any other massively computer graphics enhanced movie, then you have witnessed the unprecedented and unsurpassed graphical power of SGIs (Silicon Graphics Industries) MIPS driven IRIX machines, an UNIX flavor with especially strong graphics capabilities. These rather expensive, but both in terms of CPU and graphics power very highly ranked UNIX systems give the ability, to perform complex three dimensional simulations rather quickly to be calculated and displayed, and are still necessary for at the edge work on such projects. And IRIX has set some standards in the GUIs on UNIX too...

Apples Mac OS X: Unix for Dummies

Mac OS X Darwin kernel logo If you ever happened to know NextSTEP, an older UNIX system --- the only one, which doesn't use the X window system as underlying layer for the GUI ---, you will feel familiar with this newest UNIX flavor: technically a combination of an object oriented microkernel (Darwin, similar to the Hurd/Mach projects), BSD UNIX framework and that very same NextSTEP GUI components in a more modern version, this may become the true UNIX for everyone in the next time --- if it isn't already. While NextSTEP is renowned for it's GUI, the incompatibility to the UNIX standard X window system causes some trouble regarding usage of X based UNIX software, despite command line oriented applications are easy to port. But there is already a FreeX implementation available for Mac OS X, offering an easy and fast lane for this at least in the near future. The time to come will show, if the direction is, that more Carbon/Cocoa/Aqua native or more X server requiring ported software will appear... UNIX experts have to keep in mind, that most configurations are performed in the NextSTEP inherited NetInfo system, and not in the usually used /etc directory, where entries reside, but are mostly not functional.

The UNIX Future beckons...

Concluding this somewhat fragmentary account, I want to underline some trends, which are commonly believed to hold true in the years coming. As any such propositions can be wrong, this one makes no exception; but I'm confident, that the major parts will prove right.

The Kernel: the most advanced are the above mentioned object oriented kernels like Darwin, Mach or Hurd. They are not only easier to maintain than classical ones like the Linux or Solaris kernels, but offer a valuable advantage in a technical way: they can "hot-plug-in" new kernel components during run- and lifetime, even when the functionality wasn't prepared at kernel build time --- such things are only possible on a rather limited base with more classical kernel architectures. This may become the standard kernel model of the future.

The command and configuration layer: /etc is further the most likely place, where UNIX system configuration will take place, and the commands with the best perspective are the GNU tools, today already often not only used on GNU/Linux, but also on other UNIX flavors due to their ubiquitous presence, capabilities and uniformity.

The GUI: as mentioned above, the question is not, if GNOME or KDE (both are based on the X server and it's window system) will win, but if the X window system will keep its prevailing position. The NextSTEP/Mac OS X alternative, called now Aqua, may change things, but this is a matter not only of time, but if the Aqua system delivers enough unreachable advantages over X, to substitute it anyway.

UNIX flavor consolidation: due to the efforts of IBM (contributed the AIX JFS under the GPL), HP, Compaq and SGI (contributed the IRIX XFS under the GPL too) and some others it's most likely, that AIX, HP-UX, Tru64 and IRIX will vanish in the medium to long run and being replaced by GNU/Linux or it's successor. Then only the reference UNIX of Sun, Solaris, and Apples NextSTEP successor, Mac OS X will stay in the UNIX world as own contributions; maybe also OpenUNIX (Caldera), FreeBSD or other BSD UNIX systems will live on too. Anyway, standard efforts like the FHS (File system Hierarchy Standard), LSB (Linux Standard Base) and UNIX 98 (and its successor[s]) should reduce the differences between any different kinds of UNIX further, and altogether the management of UNIX systems will be eased up further a lot.


 

back to UNIX alliance main 

remarks etc. to: stefan.urbat@apastron.lb.shuttle.de

(URL:  http://www.lb.shuttle.de/apastron/uniForce.htm)