Last weekend I spent a couple of hours at the Orange County Auto Show, and looked at all the new cars. In the picture above, I am looking longingly at a Lamborghini Gallardo – a piece of art, as far as I am concerned and better housed in a living room than in a garage. Unfortunately, selling this type of car has suddenly become quite problematic as so many deeply-discounted used ones are coming back on the market as their owners struggle under the current economic conditions. There was also a collection of older Corvettes and Camaros on display - and they attracted just as big a crowd. And I couldn’t tell you the number of times I heard “I used to have one of those! I wished I had kept that one!”As I watched the crowds looking at the older cars, I was reminded of the things we elected to keep, and the things we just can’t wait to unload. And even for me, with the numerous cars and motorcycles I have had over the years, there’s been several that I still regret having sold. My Honda CB72 250cc motorcycle that I had in the early 70s and that I had rebuilt as a replica café-racer was definitely a “keeper”. And selling my ‘94 twin-turbo Mazda RX7, affectionately known as “the glove”, I regret it, and I will probably regret for many years to come.
When it comes to computer systems, I really haven’t developed a similar affection, or heard of anyone wishing they had held on to any particular model. Computer systems, and the technology lifecycles driving them, ensure that new systems are just so much better than anything they supersede. In the late ‘70s, while living in Sydney, I went into the warehouse of a major third-party leasing company and saw rows of idle mainframes and peripherals collecting dust. Who could have guessed, for instance, that old ‘60s era check-sorters, originally designed to channel-attach to the IBM 360 mainframe, would come back to life almost two decades later, attached to the latest iteration of a Fujitsu plug-compatible-mainframe (PCM), as happened at the Reserve Bank of Australia.
Much has changed these past thirty years. Today, systems from HP, IBM, Sun, Fujitsu, etc are all reliant on chip manufacturers. Now it’s the Intel, AMD, and IBM chip roadmaps that drive new systems into the marketplace, as each vendor tries to get to market first with technology that’s faster, more reliable, cheaper, and more importantly today – much cooler. With the race to multi-core chip technology in full swing, I just cannot imagine any user longing to return to a dual-core or even a single core engine no matter how strong the emotional ties. Systems become legacy for a reason, and hanging onto them as their components age usually isn’t a smart business decision. MTBF and MTTF suddenly cease being acronyms that are vaguely familiar!
Furthermore, I just cannot imagine a “collector’s edition” of early Itanium or Power based systems being highly sought after at auction houses. Anxious bidders will not be pushing the price of any systems built around these technologies to the same heights as we see today with art, jewelry, and cars! CIOs keep their anxieties reserved for briefings on exactly when Tukwila, Poulson - and the secretive Kittson – will begin to ship. What will follow 32-nanometer technology, that’s what they would like to know!
Before I came to the US in the late ‘80s, I had mounted into the brickwork of the office in my Sydney home the complete front panels from an IBM System 360/65 and an IBM System 370/138 (from QANTAS and from 3M I was to later find out). I had salvaged them both from a third party leasing company in exchange for a couple of cartons of beer, and my father did a terrific job of fixing them to the walls! As rustic pieces of art decorating the walls of homes and offices – could this be their only attraction?
I have been feeling a little sorry for these companies in the business of re-cycling older systems. Throughout the ‘70s and ‘80s these companies served a loyal base of customers. Should any vendors product offerings suddenly evaporate with newer models incompatible to the investments already made – then no worries, a couple of additional processors could easily be found in the warehouses of these companies. And should the economy falter – then again, no worries – old systems could be upgraded and their end-of-life temporarily ignored, as purchases of new systems are put off till better times arrive.
But today, the mix of operating systems, and much needed infrastructure software, can often negate any potential savings from staying with older systems - not to mention the problems in retaining staff capable of keeping these systems up and running. Going down this path implies you retain considerable faith in your vendor. Even for use as a development, or emergency fall-back system, this path can be risky. In a recent email exchange with one HP senior manager, we just couldn’t come up with one good reason for keeping any of the older equipment installed much past a single generation of newer technology.
Many years ago, while living in Sydney, I came very close to buying a Lamborghini Countach. I have included a picture of it here, the Countach LP400S as I still carry it around. My brother Greg and I visited the dealership many times to look at the car until finally, one
weekend, the dealer let me take it for a ride. I took it on an extensive drive around the coast roads, south of Sydney, and really bonded with the car. The trunk lid didn’t close properly, the side mirrors were useless - giving you only a view of the wheels of the vehicles around you, and you couldn’t hear the radio at all. It was perfect! But I never did buy the car, and even now every time I see a red Lamborghini, I begin to daydream about what it would have been like and whether I should look at ways to get one!And this is where all similarities end. Italian cars retain a beauty and a sense of style for decades. But enterprise computer systems never score points for beauty or style. They have to meet strict TCO and ROI figures – something completely alien when considering the purchase of any Italian car. If these cars break down, it only provides more material for bar-room discussions with the lads. If you crash them, it’s just a case of shedding tears and writing big checks! Or so it had seemed to me until this month’s issue of the TandemWorld newsletter arrived in my inbox.
I was surprised to read “the Second User market has seen a sudden and dramatic surge of inquiries for used hardware as the economy seems to be encouraging many Non-Stop users to consider extending the life of their S Series platforms. Availability of high end S Series hardware is tightening, and within 30 days we expect many upper end CPU’s to be completely unavailable.” The writer then adds “Not only is the supply tightening, but we have also seen a dramatic increase in requests for S86000, S7800, S7600, and even S7400 CPU’s lately. All of these products are also becoming scarce due to this market condition.”
The dramatic downtime in the economy has devastated many CIO’s budgets and systems are being held onto very tightly – with used equipment plugging the gaps until budgets free up. Service Level Agreements (SLA’s) remain in force and performance criteria still have to be met. Suddenly the technology and product roadmaps are no longer the full story and holding tightly onto soon-to-be-legacy equipment has nothing to do with emotions but rather, non-existent cash flow. Today’s credit markets are once again materially impacting the technology landscape. And I can only see it getting worse for many users. If the supply of older systems is drying up, as it would appear to be, then it’s a clear sign that companies are holding onto them and squeezing as much life from them as they can.
And it’s quickly developing into a sellers market, as is so often the case in difficult economic times – not everything holds its value and I am not all that sure some vendor’s equipment will hold up as strongly as others – but it’s a very good sign for anyone following the fortunes of NonStop. If CIO’s were considering this the right time to defect from NonStop, then many more systems would be available – but they aren’t. Perhaps their faith in HP and NonStop continues to be well placed after all. After so many years, could NonStop servers really have become keepers?
If a system becomes inexpensive today, wouldn’t it be a sign that perhaps their vendors may be in deep trouble? And if they start flooding the market, then doesn’t this say a lot about their user’s confidence in the platform? I may yet see a Lamborghini being sold so cheaply, as its owner cashes out quickly with no thoughts about keeping it, that temptation crosses my mind! But a good system will always be a keeper. And there’s pretty solid evidence that suggests NonStop remains a very good system well worth hanging onto during these disastrous economic times.
Comments