NonStop - A Running Commentary – February, 09

The opinions expressed here are solely those of the author
and shouldn’t be viewed as reflecting any views
of his current or past employer

I have been exchanging emails with NonStop users and vendors of late, and I am impressed with their genuine enthusiasm they have shown for NonStop. Whether it’s after  installing the latest NonStop server, as I have observed following a couple of recent BladeSystem deployments, or after porting new solution to NonStop, there’s an increasing belief that NonStop will be around for years to come.

While HP executives tell the story of how big an investment HP has made in NonStop since the acquisition and how it exceeds $1 Billion, once you begin to dig beneath the hype and look at what we have today, the roadmaps are very different to what was offered before HP arrived on the scene. And NonStop hitching a ride on the blades-train has much to do with this renewed belief in NonStop.

In the presentations that I gave last year, one of the introduction slides I used included a quote from one of my most reliable sources – the car magazine Road and Track. The columnist Dennis Simanaitis can always be relied upon to come up with something relevant for IT. Back in October, 2004 he quipped “The Stone Age didn’t end because of a lack of stones!” He then added “new technology will succeed not because it’s forced on us, but because it’s better!” And I have this feeling this is why we are seeing continued user and vendor support for the NonStop platform – NonStop on Blades is just a better solution.

As reported in the Summer 2008 issue of the 24 X 7 magazine, when NonStop on Blades was featured, HP made it very clear that “fully bladed infrastructure defines HP’s long term vision.” With this explanation as a starting point, it requires very little imagination to expect the “real estate” on the blades to be standardized as well later this year and include a finite set of interconnect, channel, and communications pipes. The price for these new blades should drop even further, as a result, and it will broaden the appeal of NonStop on Blades and attract the attention of even more IT professionals.

The continued investment in the open story and, with the deep-port of SASH that includes a combination of well-known open source infrastructure components and provides a run-time environment for Java applications on NonStop, provides an even more compelling argument in favor of NonStop on Blades. It requires even less imagination to anticipate seeing Java applications moving easily between different platforms – perhaps even automatically, based on parameters such as the requirements for availability or for data base access! Particularly should shared-blade infrastructures see the light of day as a fully supported product line.

And talking of infrastructure, last weekend I returned to Simi Valley after spending a couple of days in Las Vegas. The highway that cuts across to the West, bypassing the major freeways of Los Angeles, goes through the small township of Mojave – where commercial airlines park their excess inventory. The two sets of traffic lights in the town brought the highway to a standstill, as vacationers returning from the ski resorts of Southern California were forced to merge with the traffic from the casinos of Las Vegas. The bottleneck that these traffic lights created, allowing only a few cars at a time to pass, was just plain ugly!

This reminded me of another comment by the columnist Dennis Simanaitis, who in the September, 2008 issue of Road and Track wrote “too many cars can screw things up, even without a bottleneck …. or, said more eloquently … multi-particle interactions display a threshold of fundamental instability!” Expressed another way, throw too many different types of servers into the data center, and you risk destabilizing it.

From the time I first saw the prototype of a hybrid, shared infrastructure, BladeSystem running a mix of NonStop and Linux on industry-standard blades, as showcased at last year’s HPTF&E, I realized that it had the potential to attract even more users to NonStop. This configuration was described in more detail in the blog posting of July 16, ’80 “Specialist! Am I still needed?” where I explained that the Engineering Prototype (EP) was “not a product, as people were quick to point out, and still requiring considerable input from Product Management before it ever made it into production – it provided generalists with the opportunity to have a cluster of specialty servers delivered to them by HP and functioning right out of the box!”

There will continue to be requirements for multiple platforms, supporting different transaction mixes at different price points - not all transactions or data are of the same value to a corporation – and having them integrated and transparent, managed via a single operator interface, significantly reduces the number of “multi-particle interactions!” No issues of instability from a package like this that I can see.

There are still uncertainties surrounding the future of this shared infrastructure product as HP Product Management develops the right marketing messages, and maybe it will never be released. But I am hopeful that it will, as I see a robust future for such a hybrid product – imagine the ability to support an application as complex as that at Sabre where routines with a disproportionate number of data base inquiries, versus actual data base updates, could easily be split between NonStop and a commodity platform like Linux. Data base “tiering”, seamlessly integrated and packaged as simply as this “out of the box”, could become a major marketing message that gains traction across a broad spectrum of the IT community.

This takes me to the final quote I will include from the columnist Dennis Simanaitis when, in the October, 2007 issue of Road and Track he complained bitterly how “there are plenty of maybes, a fair amount of hard news – and even more hype!” While he was referring to the introduction of a new car, you could almost believe he was writing about the new NonStop “open” and on blades!

There is no escaping the impact the downturn in the economy is having on us all – and vendors like HP aren’t immune to what is hitting the rest of us. In a blog entry that I posted last week to the Connect web site, I asked the question of whether all HP platforms will survive. Check out the posting of February 20, 2009 “What will miss the cut?” – the good news for the NonStop community is that I talked about the possibility of OpenVMS being vulnerable in these economic times, and how NonStop could be under a lot less pressure.

And it’s not just the fate of HP platforms either, there’s much the same issue at IBM where the System i is leveraging System p hardware and could end up just being another “container” supported within their Unix offerings. Even the mainframe may have to be redesigned to better share the commodity blades available to the Unix and Linux product lines and a circumstance that would bring the mainframe much closer to looking like the HP BladeSystem we see today.

Further evolving NonStop to share commodity hardware with other commodity platforms, unchanged, will certainly help it ride out the economic bad times we are experiencing today. When it comes to critical “forks in the road” and investment decisions have to be put under the microscope, it’s going to be very hard to ignore the value proposition we are beginning to see from NonStop on Blades.

It’s an amazing testament to the foresight of the early engineering teams that, more than 30 years later and, having survived the threats from numerous companies with maybes (as good as a NonStop), the NonStop platform we deploy today remains as relevant as it ever was. Technology lifecycles generate many product lifecycles within them – and when it comes to processing mission critical applications, the NonStop product has successfully reinvented itself many times. There have been many shifts over the years with the technology pendulum swinging back and forth between centralized and distributed approaches and between in-house development and outsourced services, but the NonStop continues to be at the center of many of today’s most demanding mission critical networks.

Simply put, nothing better “out of the box” has come along that has led to the end of NonStop – and its simplicity remains a reminder to us all of how future BladeSystem packages may lead to even better integration of NonStop with other servers in the data center. And that development is neither a “maybe”, nor more “market hype”. Watching the uptake of new BladeSystems by the user community, and the enthusiasm it generates, certainly suggests even better products and solutions to come.

Richard Buckle

HP – ITUG:            Board of Directors (2000 – 2006)    Chairman (2004 & 2005)

IBM – SHARE:      Board of Directors (2007 – 2008)

Real Time View blog:           www.itug-connection.blogspot.com