Silver Anniversary For Silverlake
June 24, 2013 Timothy Prickett Morgan
It is hard to remember a time when I didn’t spend each and every day with the AS/400 and its progeny. For precisely half of my life, I have spent my days and many of my nights trying to understand the IBM midrange platform, the clients and other systems it interacts with, and watched how it has evolved over time to accommodate new technologies like client server and the Web. I have received a wonderful education about business and technology from countless midrange shops, with the only tuition fee being my own sweat equity and the coursework nothing but the good intentions of the community I serve.
It is and has always been my pleasure to serve, just like the canonical New York City coffee cups say, and I hope that you will grant me some time to be philosophical and metaphorical. The thing that I know about the system that was developed as Project Silverlake in the middle 1980s and launched as the AS/400 on June 21, 1988, is that this is how the people of this community all feel. They all want to serve the platform. The AS/400 had community long before Linux made it cool.
And I think this fact, more than any other facet of the AS/400 and its progeny, has made the difference in terms of the longevity of the platform that is the result of some extremely forward-thinking ideas about how systems software should be created to make it easier to create and maintain application software. That is a pretty significant thing to say, given all of the excellent hardware engineering that has gone into the IBM midrange platform over the past 25 years. And, indeed, you can trace these ideas back to 1979, as embodied in the System/38 that is most like the AS/400, and to 1983, when the System/36 was launched. The AS/400 is a mash up of the System/38 and System/36 platforms, retaining as many aspects of both as was technically and economically possible at the time, just like today’s Power Systems platform is a mashup of the RS/6000 and AS/400 with a bit of Linux-X86 server tossed in for good measure.
As I said last week when reporting that Hewlett-Packard has decided not to port the latest OpenVMS 8.4 release to the current “Poulson” Itanium 9500 processors from Intel and has basically sunsetted the hardware platform on the older Itanium 9300-based Integrity servers that are several years long in the tooth, it is important not to gloat. But, having said that, it is Silverlake’s 25th birthday, and it seems appropriate to keep score.
Let’s keep it simple. The major proprietary platforms that the System/3X and AS/400 competed against are largely gone. And it is important not to gloat.
For all intents and purposes, entry and midrange mainframes are just about gone from the market, excepting a few machines that IBM, Unisys, Fujitsu, and Bull still sell. The HP 3000 platform was mothballed years ago, and the DEC VAX and AlphaServer, with this latest announcement from HP, will be supported for a long time (at least until 2020 with standard support) but no one is going to write a new application for it because by 2020 new Itanium systems to run it will not have been available for sale from HP for four years. All OpenVMS apps will be in maintenance mode, as they probably have been for years. Thanks to the shock to the Itanium ecosystem caused by Oracle a little more than two years ago, when the company withdrew support of its current database, middleware, and application software on those then-impending Poulson Itaniums, Oracle has really crushed the HP-UX platform as much as any other competitive pressure from Big Blue with its AIX Unix variant, and ironically, Oracle’s shots at HP-UX and Itanium have only called into question the viability of the Sparc hardware platform and its Solaris Unix variant. And, while the Power Systems platform from IBM has utterly taken over the Unix market, it is being decimated and is one third the size, in terms of revenues, of its peak a dozen years ago.
I don’t say these things with any kind of glee. I love all server platforms, if somewhat unequally (we all have our favorites among our children, I suppose), because I am an ecosystem player and I understand that you never want a monolithic culture. It is just plain unhealthy for a variety of reasons. And that is why it is vitally important for ARM processors–the kinds we have in our tablets and cell phones–to take off on servers. ARM also has a very broad community, and while it is not open source, there are many more licensees of ARM technology and none for x86 or Power technology. Those technologies are controlled by the vendors, but ARM is a collective effort. The competitive pressure that ARM brings to all servers is healthy, just as proprietary minicomputers kept mainframes from stagnating and Unix made OS/400, OpenVMS, and MPE all evolve and adopt Unix APIs and runtimes. It was good for the IBM midrange to put the pressure on mainframes and then to feel the pressure from below from the transition to RISC processors in the late 1980s and early 1990s and then from X86 processors and the Windows and Linux operating systems in the past fifteen years or so. The infusion of various Internet technologies has been key to the long-term viability of the IBM midrange.
We only adapt and grow under the presence of an outside force, as you well know.
Just because it has been good for the platform does not necessarily mean that it has been good for all members of the AS/400 community from 25 years ago. Many of the hardware distributors, used equipment dealers, financing companies, software developers, and peripheral suppliers that I knew two and a half decades ago are not just gone, but long gone. And I will admit it right here, right now. I find that a bit disconcerting. But the motto of this company–well, the revised motto, anyway, after optimistically (some might say foolishly) trying to take on the incumbent giants of the IT publishing world with IT Jungle starting in 2001–is survive, adapt, repeat. And while that is a very hard path, it is the only one I know. And all of us who are still here, working and living in the OS/400 and IBM i community, are an expression of that mantra. And we have a slew of new vendors coming into the market, building clouds. This is important.
Instead of celebrating the AS/400 platform, as IBM and others (including us) have been doing for the past several weeks, on this Monday that is closest to the AS/400’s 25th birthday, I think that we should also celebrate each other. We came this far is the pun embedded in the name of the Italian city of Venice, established after an exodus after the fall of Rome.
It has not been easy to come this far in the AS/400 community. The amount of change that we have all had to absorb in the IT market in the past two and a half decades has been truly staggering. (It is a good thing that we all enjoy this, eh?) The technology and economics of data center computing–what still should be called data processing and transaction processing as far as I am concerned–are so changed as to be utterly unrecognizable. Honestly, when I think of myself as a young boy and look at myself today, I get the same shock, and I am sure you get the same shock, too, when you consider yourself.
The OS/400 and IBM i platform spent the first half of its life being normal and desirable only to find in the second half (thus far) that it is different from the mainstream, something that needs to be justified and defended as well as appreciated and respected for the beauty and wisdom it embodies. The platform we all earn our keep from is by no means alone.
It was hard to imagine during the dot.com boom that Unix, which comprised about half of revenue market share for servers, would ever be defending itself and outside of the mainstream, but here we are and that is absolutely happening. And as difficult as it might be to envision Windows or Linux facing a similar threat, they most certainly are, but only this time, it is more subtle. This is ever the way. But we are still there, as an IBM i community, and IBM is still supporting the platform on its latest iron and shows no inclination to change that. Unlike HP and its platforms.
The threat this time, to all operating systems and databases and middleware, is the very concept of an operating system and the practice of owning and managing a physical system, or collection of them, to run these bits of software. The new system is a ritualized, platform-layer system that, in the bitter-sweetest of ironies, is virtualized and abstracted above the iron and that presents database, runtime, and other aspects of the system as a service. As I have pointed out before, the Amazon Web Services cloud is a modern AS/400 of a sort. And one that is not compatible with the AS/400’s progeny. Others, to their credit, are trying to build cloudy services that are based on IBM i, and we celebrate them. This is hard to do. Even IBM is struggling enough with its SmartCloud that it needed to buy SoftLayer and it will probably need to buy Rackspace Hosting before someone else does. And that may still not shift any momentum away from AWS, Google App Engine/Compute Engine, and Microsoft Windows Azure.
For me, the astounding thing, given all of this change, is that we are still a community of common interest, still trying to get good work done, and still getting it done. And that was what I drank to while attending the International Super Computing conference in Leipzig, Germany, on Friday night when the Power Systems-IBM i platform entered its 25th year. Yes, she has a little silver in her temples, just like me these days, but we still have work to do. I am happy to be her sidekick, because we have skills and we can still kick ass. And that is about as good as it gets.