A year ago, McKinsey Quarterly published a special edition, Management: The next 50 year, to commemorate its 50th anniversary. To illustrate how different the world was 50 years ago, its lead article characterized the 1964 environment with three key events: the Summer Olympics in Tokyo, which was the first global telecast via satellite; the last of the baby-boomers were born that year; and IBM announced the new System/360 family of mainframes. I was reminded of that third, distant event by IBM’s recent announcement of LinuxONE, its new Linux-only mainframes.
In 1964 I was a second year college student at the University of Chicago. I was also working part-time at the university’s 2-year old computation center, and still remember attending a presentation on the new System 360 by a visiting IBM executive. I later went on to graduate school in physics at the university, where I made extensive use of S/360 computers for my thesis research. In IBM, which I joined after finishing my studies in 1970, major portions of my 37 year career involved mainframe architecture and strategy.
But it wasn’t only my long association with mainframes that this announcement reminded me of. The Linux side of the announcement also brought to mind some more recent memories.
The Linux kernel was first released by Linus Torvalds in 1991. Linux started to take off in the mid-1990s, primarily in the supercomputing community, which saw it as a way of replacing their expensive machines with clusters of Linux-based commodity servers. Later that decade, Linux was also widely used in a variety of Internet applications, including web servers and firewall systems.
By 1999, Linux interest was increasing in the more general IT marketplace. That summer, IBM commissioned a study to help define the company’s overall Linux strategy. After three month, the study strongly recommended that IBM embrace Linux across all its product lines, that it should partner with the Linux community and contribute to its development, and that it should establish a company-wide organization to coordinate its various Linux activities.
In January of 2000 IBM announced the formation of the new Linux organization, which I was appointed to lead. A few weeks later, I gave a keynote presentation at the LinuxWorld conference in New York to explain IBM’s decision to embrace Linux across the whole company, including mainframes, - perhaps the most surprising part of our Linux strategy. But it turned out that the mainframe’s ability to support many virtual machines within one system was particularly suitable to support many Linux applications within one physical mainframe. Before long, Linux on mainframes became well accepted in the marketplace.
Let me briefly summarize what IBM announced a few weeks ago:
- LinuxONE, - two Linux-only mainframes, one aimed at large enterprises and the other at mid-size companies. In my opinion, the most important part of the announcement is LinuxONE’s support of the advanced encryption features that have been built into the latest mainframes models, enabling them to support up to 8,000 highly secure Linux systems in one machine.
- Seamless support of open source tools and software, including Apache Spark, Node.js, MongoDB, MariaDB, PostgreSQL, Chef and Docker. These work on the mainframe just as they do with other platforms and thus require no special skills.
- The LinuxONE Developer Clouds, a kind of virtual R&D engine to help developers create, port, test and benchmark LinuxOne applications.
- The single largest contribution of mainframe code to the open source community, including advanced predictive analytics that constantly monitor for unusual system behavior to help prevent serious system problems. These contributions can be used to build similar sense-and-respond resiliency capabilities on other systems.
“Altogether IBM is trying to make the mainframe a more relevant computer for a broader variety of businesses, and put it in the center of trends like big data and mobile,” notes this article on the LinuxONE announcement. “The mainframe is one of those products that’s been declared dead a dozen times over… However mainframes have remained the preferred computer of choice for big companies that need to crank out huge numbers of transactions.” Especially, I would add, for those transactions that require the highest degrees of security, availability and data privacy.
Having lived through the near demise of mainframes in the early 1990s, - which would have inevitably led to IBM’s own demise, - their ability to have survived after all these years is truly impressive. Many once great computer products, - and the companies that developed them, - are no longer around. Few computer families can trace their vintage to the 1980s, let alone the 1960s. There is something pretty unique about the mainframe being not only alive but doing quite well after all these years.
Why is the mainframe still around, 51 years after first being introduced? What enables it to keep reinventing itself while embracing the latest technologies, including cloud computing, big data and analytics, connections to all kinds of mobile devices and a wide variety of open source software and tools? In a world where product life-cycles are measured in web years, what can we learn from the mainframe’s rather unique longevity?
These are questions I’ve long been thinking about. Over the years, many have predicted that the end was near. Some were competitors hoping to replace mainframes with their own products. Others just couldn’t imagine that anything this old could still have value in such a fast changing industry. In March of 1991, for example, VC and journalist Stewart Alsop famously wrote in InfoWorld: “I predict that the last mainframe will be unplugged on March 15, 1996.” A decade later, with mainframes still alive and going strong, Alsop metaphorically ate his words.
Last year, as the mainframe marked its 50th anniversary, The Register published an article which asked in its title: Why won’t you DIE? IBM's S/360 and its legacy at 50. “Fifty years after the first S/360 was announced and 30 years after the rise of distributed systems that were supposed to replace them, the mainframe is smaller in market share, but its principles are being embraced once again.” Let’s take a look at some of these principles.
To a large extent, the mainframe’s longevity is a result of two major architectural innovations first introduced with S/360. The first was the notion of a family of computers, from low to high performance, all based on the same instruction set which allowed customers to upgrade to larger systems as well as to future models without having to rewrite their applications. The second was OS/360, a common operating system that supported the various members of the S/360 family except for the smaller ones which ran a subset with more limited capabilities. IBM’s z/Architecture and z/OS are direct descendants of the original S/360 and OS/360.
These were radical ideas back then. They ushered a whole new way of thinking about computer systems. Prior to S/360, each new computer model was built from scratch. So was its operating system. Software designed to run on one machine would not work on other machines. Today, these architecture and business ideas are well accepted and embodied in the concept of platforms, which make it possible to implement an architectural design across different technologies and models, as well as enabling the formation of an ecosystem of complementary products, applications and services.
These platform innovations have enabled mainframes to incorporate major advances in technology and keep evolving over the years. This was the case with the transition to microprocessors and the parallel sysplex architecture, without which IBM would not have survived its near-death experience in the early 1990s. A few years later, mainframes embraced TCP/IP and just about all Internet standards, integrating seamlessly with the Internet and World Wide Web, and enabling its customers to leverage their existing transaction and data base applications in all kinds of new e-business solutions. Then came zLinux, which made it possible to easily port just about any Linux application to mainframes.
The IBM-Linux story also had to overcome a number of challenges. Initially, IBM’s 2000 Linux announcement got a mixed reception. Many welcomed IBM’s strong support of Linux and open source in general. But Linux was still not all that well known in the commercial marketplace. Some viewed Linux as just another operating system developed by a bunch of hackers, and were totally perplexed that IBM was so aggressively supporting an initiative that, in their opinion, was so removed from the IT mainstream. Gaining their support took considerable time and effort.
A few companies viewed Linux as a competitive threat and strongly attacked not only IBM, but also companies that used Linux as well as the overall Linux community. One famous CEO described Linux as akin to communism and cancer. The attacks against Linux and the companies that embraced it continued for years, including the SCO Group’s 2003 $5 billion lawsuit against IBM.
Despite it all, mainframes and Linux have moved forward, continuing to play major roles in our 21st century digital infrastructures. “Fifteen years ago IBM surprised the industry by putting Linux on the mainframe, and today more than a third of IBM mainframe clients are running Linux,” said an IBM senior executive in the LinuxONE announcement. Given that mainframes and Linux have been such major parts of my own career, it’s particularly gratifying to see how well they’re both doing after all these years.
Comments