Sunday, January 3, 2010

Tandem – the next 35 years?

Over the holiday season I didn’t spend as much time pulling bottles of wine from the cellar as in previous years, yet sadly, I have to admit, the inventory is beginning to thin out. On the other hand, what has been growing has been my collection of model cars and motorcycles and now, spread across the tasting table, they capture quite a bit of history – the picture above includes most of what is on display.

Obviously, there are several model Corvettes in the collection. Among the models are F1 cars – the Ferrari of Schumacher and the Jaguar of Weber – as well as a Yamaha R1 sports bike that’s a constant reminder of the fickleness of youth! Missing from the display is the one car I thought would always find a place in my collection – the mighty Porsche Turbo!

When I returned to Australia in 1977 I took up an assignment with a small software distributor that represented The Computer Software Company (TCSC) of Richmond, Virginia. However, in those early days I was in desperate need of technical support so TCSC assigned a young IBM systems programmer, Brian, to help me out. Business began to go well and I thought an upgrade of my company car may be in order and I began to think the time was right to buy a Porsche Turbo.

Dreaming big was rarely an issue for me, and I had become extremely passionate about Porsches. One evening, on returning home, I found Brian waiting for me - all smiles and looking excited, with a bunch of papers in his hand. “Sign these papers,” he prodded me. “There’s a surprise in the garage for you!” The papers I quickly realized was a car leasing agreement, and Brian cagily hid most of the pages from my view but just once I happened to see a reference to a Porsche Turbo. With the paperwork completed the time had come to head for the and, with the door opened, there in the middle of the garage was a Porsche Turbo – but in miniature. It was a Matchbox model!

“As always you were a ‘voracious’ car enthusiast,” Brian reminded me this week. “Russ, one of your neighbors, had boys with Matchbox model cars, so we ‘borrowed’ the Porsche and placed it in the garage.” Recalling more of the incident, Brian then added “Russ was very good at leading up to the idea that the financing was not really an issue … but your face dropped when you saw the miniature!”

The Porsche Turbo remains the pinnacle of auto engineering, and my passion for the car remains to this day. So you can imagine my surprise when I began reading this month’s car magazines only to learn that the Porsche Turbo is 35 years old! Recognizable worldwide, the Porsche Turbo continues to promise unbelievable performance in the hands of those passionate enough to wring the best from such a supercar!

A new engine? The first upgrade of the engine “form factor” and, a first for this latest iteration, an engine that shares the same construction as the standard 911! With the same direct-fuel-injection system as found in all 911s, this “new engine” is simpler and lighter than any previous engine, with commodity components prevalent everywhere! As I read the specifications I almost forgot I was looking at 35 years of Porsche history and not at 35 years of Tandem history!

Pinnacle? Passion? Promise? The emotions they generate are very similar for both product lines.

Cars have been featured in my past two blog postings on Tandem’s 35th anniversary with references to the late Graham Hill in the posting “Happy 35th, Tandem!” on May 28th, ’09 and to Corvettes in the posting “Is there gas in the tank?” on October 16th, ’09. In both postings I also included a number of quotes from current and former leaders of Tandem.

The two comments however that continue to resonate with me were those of Martin Fink and Winston Prather. “From the introduction of NonStop until today, NonStop has represented the pinnacle of high availability,” Martin had told me. And then, as I exchanged emails with Winston, he talked openly of how he hadn’t “fully understood then was the incredible passion and commitment this team has to our customers, partners and each other!”

In the November / December 2009 issue of the Connection magazine, Winston wrote in his column of how “from our origins as the world’s first fault tolerant server, through mergers and acquisitions, the evolution of IT and its role, processor and form factor changes, the original promise of NonStop has never changed.” He then closed his column with “we are as committed to delivering on this promise today, as we were 35 years ago.”

There has always been passion in the NonStop community. And the NonStop server has remained at the very pinnacle of fault tolerant computing. But what will the next 35 years promise for the NonStop faithful? Will support for NonStop continue to play a major role in HP’s server strategy?

In a blog posting of February, 12, ’08 “’My Wish’ for NS Blades” I described three wished I had for NonStop as it began supporting Blades. My first wish was on seeing a “shared infrastructure” Blades package delivered, while my second wish was to see a hypervisor developed such that NonStop could operate as a guest OS, and where the fault tolerance of NonStop wasn’t compromised in any way. As for my third wish, I was looking to see if a case could be made to support a NonStop that could float and “adjust” automatically whenever business-critical transactions arrived that required the levels of fault tolerance NonStop delivers.

I have been talking of this to many audiences over the past two years, but I was somewhat taken back when I read the latest white paper from The Standish Group “Megaplex – and Odyssey of Innovation.” Jim Johnson and his analysts at Standish talk about the possible emergence of “Megaplex, a collection of server blades acting together as a single system using multiple types of operating systems, databases, and other computer resources.” Standish recalls Tandem’s history and describes a future role for NonStop that is far removed from anything that has found its way onto HP’s roadmaps to date.

Standish was looking at the characteristics of NonStop that would help support the NonStop platform for another 35 years. In perhaps the most remarkable observation in the whole white paper, they propose how “The Standish Group sees the death of operating systems and the advent of operating environments … the operating system of tomorrow will be a lightweight hypervisor that calls on a set of common services,” starts Jim. “In the Megaplex fabric these hypervisors will each have their own persona, such as VMS or the NonStop Kernel, to facilitate applications services,” he then explains before adding “many of the NonStop Kernel functions, such as fault tolerance, will be available to these other personas to increase their functional capabilities.”

When I posted my own observations to the blog in February, ’08 I predicted that, at some time well into the future, NonStop would be available to process any transaction whose attributes qualified it for processing on NonStop. With Megaplex, Standish take NonStop many steps further and suggest that the NonStop kernel, as well as being able to run natively, would pass on critical capabilities that had always been unique to NonStop to other HP operating systems.

I promised myself a Porsche Turbo one day – and perhaps after all these years, it will end up being just a model after all! And the similarity between the history of the Porsche Turbo and the Tandem computer is remarkable. Thirty five years on, they both remain the pinnacle of engineering excellence and performance. And their supporters are enthusiastic and passionate about both platforms. It may be hard to imagine all of what Standish predicts actually coming to fruition, but it would be hard to bet against the folks still working on NonStop.

Jimmy Treybig has always been extremely proud of the people who worked on Tandem. In the post to the blog of May 28, ’09 I included Jimmy’s final email remark to me “I was amazed at the capabilities and drive of our people and very proud of how they met our very aggressive goals while we had great fun!” And the passion for NonStop remains today as strong as it ever has so I will not rule out anything from happening with NonStop in the future.

I don’t think anyone in the NonStop community senses that the commitment to NonStop, within the NonStop development organization, has lessened. But the effort needed to fulfill my own predictions, let alone those of Standish, will require significant investment from HP. But there’s always hope – and in his closing remarks in the most recent issue of the Connection magazine, Winston did state that “we are committed to delivering on the promise today, as we were 35 years ago. With our modern infrastructure that delivers uncommon value through common standards, the future is indeed bright.

Pinnacle! Passion! Promise! No, Porsche doesn’t have a lock on these emotions after all!


TB said...

Richard, very good "catch" in finding the parallels between Porsche and NonStop.

I think there is one major difference though: Porsche is sold in a market ("luxury cars") which still has significant growth potential (I guess mostly China, India and the like these days).

On the contrary, NonStop has been sold primarily as "the always up platform". That market is limited in size - so unless NonStop goes "beyond availability" there is no major growths potential.

Everyone is talking about "virtualization" and "cloud computing" these days. For instance, there is an interesting Gartner talk on "XTP" (eXtreme Transaction Processing) which I unfortunately cannot share in the public. In a nutshell, it says OLTP will move into the cloud where the cloud will be partially "private" (you own and run it) and "public" (someone else runs it for you).

While that Gartner vision is certainly compelling they see it happen through the software stack and I think this is bound to fail (we shall see in the next five years...).

What is sad about this: If you look at the actual requirements of what Gartner calls XTP NONSTOP HAS IT ALL ALREADY (well, nearly). I think it would be much easier to move from NonStop to XTP ("through the hardware/OS layer") than from VMware, Java XXX ("through the software stack above the OS") as Gartner sees happening.

Unless Gartner, the "big HP" and other major players agree with me here (and I doubt either will) NonStop will play no role in "the future of the cloud" :-(

Sorry to end on a sad note, but I don't have high hopes for "the NonStop platform beyond availability" right now. I would be glad if other readers would disagree with me though...

Kind Regards,

Thomas Burg

Anonymous said...

While bits of NonStop functionality may be remain in software form, lockstep, itself, has already been eliminated from the more recent offerings, and, without that, is it really still 'Tandem?'

Online Transcendence said...

Richard, I think the article referenced here expresses my feelings about HP's position in the marketplace:

I can't agree with this more... HP is truly positioned to leverage the massive intellectual property assets at its disposal, and it has an interesting angle, to become the 'soup to nuts' infrastructure provider:;jsessionid=FWKED0RW25CKDQE1GHPSKH4ATMY32JVN?articleID=220900544

I never thought I would say this, but the key here is marketing, and to some extent, branding. HP flat-out has not successfully commanded the kind of mind-share that IBM seems to. They need to actively promote their 'thought-leadershippiness' to the marketplace, by pointing out the depth and breadth of their enterprise customer base (with the NonStop position in payments/financials, telecom, etc.).