SUSE employees discuss the origins of enterprise Linux, the open source development model and how SUSE evolved into a commercial operating system.
What role did SUSE play in the growth of Linux and the open source ecosystem? How did SUSE and other Linux-based operating systems evolve into the enterprise platforms they are today? Here's what SUSE employees had to say about Linux history in a recent interview.
To help mark the anniversary of Linus Torvalds's release of Linux twenty-five years ago, I interviewed Meiki Chabowski, SUSE Documentation, and Markus Feilner, Strategist & Documentation Team Lead. Their answers, printed below, provide interesting perspective not only on the history of SUSE, but also of Linux and open source as a whole.
Q. From your perspective, what were the one or two biggest changes in the Linux ecosystem between the early 1990s and today?
Chabowski: In 1999/2000 a lot of big enterprises started to be massively interested in Linux. Just looking at my personal experience from SUSE, during that timeframe, we started to have the first real technical cooperation or partnership with IBM, SAP, Oracle, and other "big players" in the industry. These partnerships did "change the game" for Linux.
To give you more details, from a SUSE point of view: SUSE’s partnership with IBM, forged in 1999, led to a number of forward-thinking projects. For example, SUSE worked with IBM and Marist College to port available Linux code to the mainframe. A year later, SUSE was only Linux operating system for IBM mainframes that was enterprise ready, fully supported and commercially available.
Feilner: I remember clearly when a friend of mine came to my student office at home, where I had a 386 running OS2 and DOS 6.0, both with Windows 3.11. At work and at University, we had a SUN Sparc Cluster and I had been looking for the power of Unix on my PC at home. My friend said he could make that possible. He dropped by for a night of installing some 50 or so floppy disks -- 3.5 inch. It left me with a shell, no graphical UI to click and more questions than disks on my pile. What was it? SUSE 1.0, in late 1994. I know that because we were all waiting for Windows 95 or Windows 9.x, as we called it, since the year of its release had been postponed several times.
Today, we install Linux in less than ten minutes, we deploy it automagically on several systems, architectures and virtual machines. We do not use disk, "floppies," CD's or DVD’s anymore. Even image files seem to be somewhat out of vogue. If you watch modern SaltStack or OpenStack, clouds are dumping their config onto a standard VM.
We were "nerds" then, outsiders, off the beaten track, and today we are mainstream. Linux is everywhere and there are installations going on every minute. And we are legion.
Why do you believe Linux ended up massively popular, while better-funded free Unix-like kernels (such as GNU and BSD) gained much less traction?
Chabowski: The Linux kernel was written completely from scratch, and shared no code with any existing Unix. Thus it was not roped into the licensing quarreling that hindered Unix growth during the so-called Unix Wars. Free Unix-like variants somehow could not get rid of the “aftertaste” of these struggles.
In addition, the Linux community grew quickly (frustration with licensed Unix derivatives was high, the time somehow was just ripe), and the community was strong since the beginning.
Feilner: Somehow, the better development model has prevailed. Speaking of GNU and BSD, the differences may be little, but they may be crucial. Probably many people were deterred by BSD's economic implications, others may have had troubles with strange personalities like [Richard] Stallman or others -- who knows. Maybe the Linux world had just enough of both, but never too much of that. We had support from GNU and vivid exchange with all the BSDs, but at the same time the Linux world never was as conservative in its choices and was always ready to take an extra step forward.
What has surprised you most about Linux during your time working with it?
Chabowski: The philosophy of "open and free source code" impressed me when I heard about it. But what surprised me most and still does fascinate me today about Linux is that its model of joint and worldwide development worked and still works.
Linux was born in 1991, and since then it has been ported to more hardware platforms than any other operating system. Today, thanks to the spread of Android smartphones, Linux has the largest installed base of all general-purpose operating systems worldwide. Linux is also the leading operating system on servers of any sort and of 99.4 percent of the top500 supercomputers. And you find embedded Linux in a huge number of devices and machines -- built into cars, network routers, facility automation controls, entertainment equipment, medical equipment such as X-rays -- and people do not even know about it. Linux is everywhere -- this is amazing and surprises me over and over again.
Feilner: I'm from a world where we were told that to make it big, good planning and preparation is necessary. Linux told us that this is wrong -- see the Cathedral and the Bazaar. For bigger projects - in some regards the Linux project is the biggest one mankind has ever seen -- the open development model is mandatory. Proprietary ways are bound to fail. We had a hard time believing that in the 90’s, but somehow many of us did, and it still feels awkward today to see that this assumption turned out right. Apart from this, the general and broad industry support for our development model is both a surprise and overwhelming.
What was the biggest challenge for enterprises using Linux 25 days ago, and how has that changed today?
Chabowski: When Linux was still in its infancy, companies that were using it had to fight against challenges such as "missing" usability, lack of support, immaturity of products, lack of skills and lack of applications.
None of these issues exist anymore. On the contrary, usability does not differ from other operating systems, support is a crucial part of the business model, the products are highly sophisticated and well developed, Linux is more stable, secure and provides better performance than many other operating systems, and new generations of IT specialists grow up with Linux. Only the lack of applications in some very limited and very specific areas may be a last remaining challenge from the old times.
Which Linux milestone has meant the most to you and the industry?
Chabowski: (One of) The most important milestone(s) for me (and for my job) – and I think for Linux in general -- was the “invention” of enterprise Linux back in 1999/2000. This was when IBM, Marist College and SUSE began working on a version of Linux for the mainframe.
By the late 1990s IBM was engaging with the open source community on multiple fronts. Initially open source was a challenging issue in many areas at IBM. The idea of any non-native operating system running on an IBM mainframe was so controversial that an internal project to port Linux for the s/390, based in Boeblingen, Germany, was carefully kept secret from the rest of the company. That began to change in 1998/99. When the Boeblingen IBM team made the project public and looked for a Linux distributor to join forces, SUSE was the first established distributor to respond, despite the fact that many skeptics were asking, "Who needs Linux on a mainframe? It is a stillbirth."
When SUSE started to talk with IBM about Linux for the mainframe, this market was very mature and conservative. Customers expected to get the highest reliability, availability and serviceability not only from the hardware, but also from the operating system. Thus, when introducing and establishing a product like SUSE Linux Enterprise Server for System z in such a highly demanding market, it was not enough to have outstanding technology. The challenge we faced was that SUSE at the time was being sold primarily to private end-users, but not to businesses with clear requirements for a stable and maintained business. The most challenging aspect in bringing Linux to the mainframe was establishing the right level of confidence with the customers. By working with IBM on getting Linux onto the mainframe, and by learning more about the mainframe customers’ needs and requirements, we also learned how to change our previous open source software development and distribution model into an enterprise-ready business model that included maintenance, certifications and support as insurance for a customer's IT environment. We were "obliged" to introduce the very first enterprise Linux server subscription model worldwide, which today is the basis and foundation of our entire business concept and our worldwide success.
In short, Linux for the mainframe involved creating new processes and infrastructure at SUSE, a new business model, 24x7 worldwide support, synchronized IBM/SUSE Level 3 support processes, ISV and IHV certification, etc. In 2000, SUSE Linux Enterprise Server for s/390, the first true enterprise-class Linux offering, was born. And all other commercial Linux enterprise distributions adopted the new “enterprise-class” business model.
Feilner: IBM's announcement to invest 1B$ in Linux back in 2000. That definitely showed the world: Here's something EVERY company, no matter how small, no matter how big, can use in production.
SUSE's roots are among the oldest in the modern Linux ecosystem. Why do you think the distributions that gave rise to what is today SUSE endured for more than two decades, while so many other distributions (Mandrake/Mandriva, Corel, Xandros, for example) rose and fell?
Chabowski: I think there are several reasons for it. SUSE was the first Linux distribution in the market and could gain popularity quickly. Our early technical partnerships with "big players" in the IT industry also played an important part. We were always at the forefront of technical development (e.g., first Linux distro on 64 bit systems, first to support Xen and KVM virtualization, etc.). And we are recognized for listening to and caring for our customers. This proven reputation as an open-minded partner you can directly access, discuss, address your requirements with (this is also where our tagline “We adapt – You succeed” comes from) in my view was and is crucial for our success. We are building up strong relationships with our partners and our customers – because they ARE important to us.
Feilner: SUSE has prevailed because the product has been appreciated by users. SUSE has had its ups and downs and has been bought three times because the buyers wanted to use the exceptional SUSE potential and combine it with their marketing. In fact, since we are still here and better than ever, one might consider the whole process a huge success, even though it didn't always feel like one. And what made me change from journalism (I was deputy editor-in-chief at Linux Magazine) to a distributor were the innovative ideas and the start-up-like character of this company. There are many stories about what SUSE has invented or brought on the way, and others took the flowers. XEN, KVM, x86_64, z System, SAP and SAP HANA, the list might be very long, all these were projects were SUSE was first and helped make the innovative new stuff usable in enterprise. The only mistake SUSE made was: we didn't talk much about it. That's why they hired me: to do good things and talk about it.