Tuesday, November 30, 2010

Nothing seems to last ...

It was only a few weeks ago when we took our car to the track, for the last time this year. And it was a wet weekend, raining so hard on Saturday that all events were cancelled. It was still a gamble as Sunday morning rolled around as the rain continued to come down, but intrepid drivers were not prepared to give up on the entire weekend.

The picture here is of the Corvette on grid, rain easing up just a little, but with me a tad anxious all the same. The season had started with a wet weekend and so it seemed appropriate to end on a similar note.

Ending the year at the fabulous (for Corvettes) and very wet Auto Club Speedway, a combination of an infield road course with the wide open banked oval of the NASCAR circuit (and hence the label “roval”) seemed appropriate, if not a little depressing.

When there’s water on a track however, all drivers take to the circuit with a measured sense of self preservation. No matter the kind of car on track, there’s not a driver who wants to return to pit lane with a car that’s badly bent out of shape!

Wet weather driving does encourage experimentation and a little innovation – very quickly the racing line becomes visible as the only dry section on the circuit. However, it may not be the best place to drive and early laps see drivers checking out alternate lines around the circuit and experimenting with braking zones and turn-in points. There was even a brief period where we had a dry, extremely clean, track and although it didn’t last long, I managed to eke out a couple of good laps.

In the latest issue of Time magazine that I picked up a few days ago, Richard Stengel, the magazine’s Managing Editor, wrote in his column “in our networked world, nothing ever goes away, but nothing seems to last very long either.” Certainly, applicable to the weather I faced this weekend.

Stengel then goes on to suggest “information these days is a commodity; understanding is scarce.” All of us trackside could see that it was wet but it was only a handful of wily drivers who recognized where to drive to avoid mishaps. And it can be said that all of us can get our hands on reams of publications about HP and about NonStop, but do we truly understand where NonStop is headed?

The Tandem architecture is now 35 years old and most of us have warm memories from the great times we enjoyed during the early days. Yet, to paraphrase Stengel, particularly of late as NonStop rides the Intel “curve,” nothing seems to last very long and yet, nothing ever goes away!

It was over two years ago when I posted “‘My Wish’ for NS Blades” to this blog and developed a Powerpoint slide-show around its core messages that I presented at numerous NonStop user community events. As a writer ever-willing to provide opinions, it’s a sobering thought to return to forecasts made this long ago, but before providing any further predictions, it’s none the less an important first step to take. After all, how accurate were these prior predictions?

In that post of February 12th, 2008 I suggested that my wish list was comprised of just three items - HP delivers in its message of Shared Infrastructure blades, HP commits to virtualization and provides a Hypervisor that supports NonStop as a guest operating system, and that HP acknowledges that it will also need to support middleware that interrogates incoming transactions, directing mission critical transactions to NonStop, important informational but not quite mission-critical to a Unix or Linux, and voluminous inquiries to Windows.

The first item was a request of HP BCS to deliver on the promise made in the slideware Martin Fink first unveiled as the “Shared Infrastructure Blades” package. This is where any mix of NonStop, HP-UX, Linux, and Windows Server OS’s will be supported. Shortly afterwards, the wraps came off a blades chassis populated with a mix of NonStop and Linux blades, and participants at the HPTF event that year could see demonstrations of mixed workloads running on this system hybrid.

Looked good and attracted the attention of a number of vendors but NonStop customers were a little uncertain as to what to make of it and gave little feedback to HP, and so we have not heard much more about this first attempt at hybrid.

But followers of the IBM mainframe can now point to the latest mainframe offering where zOS and Linux (with other OS images to follow, it is believed) sharing a common system bus with almost zero latency between intersystem calls! With all the talk of hybrid clusters and cloud computing, industry standard chassis populated with commodity blades running any combination of OS images is a goal being pursued by many vendors.

As for virtualization and a hypervisor, like one from VMWare, is not something I’m as aggressively proposing. It’s a bit like trying to determine what tires you need to run on a wet track – and be careful what you ask for! As so many in the community highlighted for me, there would be a cost to the level of availability offered and NonStop would lose all visibility to the underlying hardware – something very akin to the dicey situation that can develop when on a wet track with the wrong tires.

However, even as I admitted at the time of how I was going out on a limb with my last wish suggesting that with the mixed OS hybrids there would be a need for middleware products that supported real time interrogation of the incoming transactions and would direct mission critical transactions to NonStop, important informational but not quite mission-critical to a Unix or Linux, and voluminous inquiries to Windows?

Even as I see the hybrid model of my first wish building up a head of steam, this third wish – a kind of mutant mixed workload measurement and management system – remains a valid concern. Looking back, I continue to stick with two out of my three wishes which, in some small way, isn’t all that bad of a result.

Watching the sports headlines this week, TCU university, like so many other American collages of late, is switching from one conference association to another. Electing to participate with other collages with greater potential to play in major tournaments, TCU was leaving behind a program that had seen it join the elite of American football teams.

However, as well as it performed on the playing field, it was still having a tough time getting the attention of the tournament organizers and was being overlooked as potentially, the best football program in the country. "If you don't dream, you're living in a memory," Del Conte, TCU’s athletic director, said. Conte than proposed "who wants to live in a memory? Every single time we have an opportunity to think about where we're going to go (we) dare to be great academically and athletically!”

And nothing could be closer to the truth when it comes to looking at frequently overlooked NonStop – yes, memories of 35 years of Tandem architecture remain with us to this day but for me, it’s all about the future. Readers may have missed some commentary I provided in NonStop – A Running Commentary in the October issue of the eNewsletter, Tandemworld.Net and the slight variation I made on my earlier forecasts.

Gone is the pursuit of a hypervisor capable of supporting NonStop, and the availability of hybrid clusters in a box is now something I sense solutions providers will be the parties first to embrace and utilize in more creative fashion. And along with the hybrids, the need for hybrid workload management hasn’t lessened in the least, and so that remains a consideration.

New are the observations of a NonStop server becoming a smart controller! And of whether NonStop Enterprise Division (NED) will develop any of this at all, becoming a clearing house, or distribution point, for an arbitrary collection of third party infrastructure and middleware!

Will product roadmaps become nothing more than templates provided as guides and reference points to an ecosystem of ISVs investing their own nickels and dimes to populate?

My previous predictions first posted in early 2008 are all but memories for most of my readers but tallying up the scorecard and getting a two thirds pass rate I will take any time. All that I need now is one of these last two observations to prove to be correct to retain that two thirds pass rate.

Having said that, these latest projected events could prove to be extremely controversial among some users and trigger many more questions in the months ahead. On the other hand, and to paraphrase TCU’s Conte, too long with nothing but memories and it’s hard to start dreaming! And we would all be the poorer if no one dreams of future NonStop deployments!


No comments: