
But first, when I began blogging I had just finished five years of association with the ITUG Board, including two years as its Chairman (2004 – 2005). Probably not as widely known was that 2005, holding down the very honorary title of Past Chairman, and as I was working for ACI during a time of transition, a volunteer was needed to participate in the IBM mainframe user group, SHARE. For two years I served on the SHARE Board (2006 – 2007) and it was with consent from the SHARE Board that I began to blog to the NonStop community. Little did anyone at the time consider that a new career path could develop from blogging, but today I enjoy the support of a very tolerant client base even as I continue to present NonStop from a slightly different point of view!
The picture above was taken of me on the Santa Monica pier around the time I began to blog and I chose this picture as, given the location among the rollercoasters and carousels of the pier, it’s a reminder of just how many wild gyrations there has been over the past five years. The Apple iPhone had just started to ship, but the iPad was still a long way off. LinkedIn (2002), Facebook (2004) and Twitter (2006) were very much in their infancy and few understood what social networking and the social media were to unleash upon the community. The NASA Space Shuttle was still flying and ITUG was still holding events. And NonStop, running on Blades, was still a long way off!
My opening question to Martin Fink was about what he considered were among the highlights of the past five years when it came to NonStop and the NonStop community. “There was a pretty big transformation for NonStop when HP went to Blades and this included NonStop,” Fink opened with. He then explained to me just how much work went into that project and just how hard it was to get everything to work. “But now we are happy with what we did. On one hand it was radical (yet) on the other hand, the transition went smoothly (within the customer base).”
However, there was also another transition that occurred according to Fink, and that was moving from Tandem, the hardware vendor, to NonStop, the software vendor. “Tandem identified itself as a hardware company and customers perceived its value was with hardware not available in any other way,” Fink explained. “So, we went through a pretty radical transformation to where NonStop is a software play; it’s where the customer sees the value today. And now we have new customers on board as well as new vendors.” Today the only unique hardware component remains ServerNet but even here, while Fink expressed caution, he acknowledged “but we do want to get to that place where we are a pure software company so ultimately, yes, it may be replaced by InfiniBand but there’s lots of work to do.”
This discussion on the
transformation of NonStop into a pure software play and on the importance
blades are playing (with the potential for blades to become even more uniform,
more commodity based, in time) led to revisiting the Converged Infrastructure
program. “When you look at how the different groups in HP had been working –
Storage, Networking, Servers – then we have come a long way and yes, you can
see from the respective product roadmaps the results from having pursued a
converged infrastructure,” Fink explained before changing gears and catching me
a little by surprise.
“You get a variety of vectors that represent different thought processes and now we are recognizing the human issue; (within IT, there’s still) domain specialists - my server groups, that is, my Unix, Windows, etc. groups, my security, networking and storage groups, and so forth. How do I realign my human capital behind a Converged Infrastructure given how today we have many domain specialists? How is human capital to be trained?” Fink did suggest however, that “with clouds and virtualization then these (practices) are causing different groups to merge.”
Clouds? Virtualization? And even, Big Data? “Another complicated vector thrown into all of this (merging of domain specialists) and where NonStop plays – how we leverage NonStop expertise – is Big Data. People are talking about how they run big databases and with what they are doing is talking-up this idea of Massively Parallel Processing (MPP). And users will tell me ‘I probably need a columnar database (which NonStop isn’t) that’s on a MPP but then, wouldn’t it be good to bring in transactional data – a mixed columnar data store with transactional data’” Fink remarked before adding “but we have been running MPP engines that are some of the biggest the world has ever seen. There’s no better engine than NonStop!” While avoiding addressing any specific NonStop involved projects, and clearly rejecting any notion about NonStop being legacy, Fink did say that there were teams working on how best to leverage the intellectual property (IP) of NonStop with respect to Big Data.
As for Clouds, the projections were a little more cut-and-dried. “Most clouds today feature a lot of scale-out ‘pizza boxes’ and as long as you can partition the data, it all works,” Fink began before clarifying that there’s not any immediate future for NonStop inside the more popular cloud options being marketed. However, when it comes to NonStop, “being on the edge (of the cloud as) the resilient cloud presenter,” Fink then added, may prove to be an option for users already familiar with (and invested in) NonStop. But then pragmatically, with clouds that exist today we see just a further iteration in the long-lived cycle of technology trends with what we are see being just another phase as users continue to pursue less-expensive processing options.
Looking ahead to what the NonStop community can expect to see over the course of the next two to three years, Fink was even more pragmatic, saying “it’s all about Poulson. And looking at what it does for our users. It (provides) another performance leap and what we shouldn’t ignore is that it’s not just another chip upgrade but the first major architectural revision to Itanium since (Itanium 2) McKinley delivered in 2002.” And according to Intel, Poulson “will use a new microarchitecture, with a more advanced form of multi-threading that uses as many as four threads, to improve performance for single threaded and multi-threaded workloads.” Or, as Fink reiterated “it’s everything we see today, plus more (that we can use) to deliver even more for the NonStop user – it will be the high point for NonStop in the coming years.”
The image at the top of this post of me standing in front of an amusement park featuring wild rides perhaps isn’t altogether a fair reflection on NonStop. Many within the NonStop development organization remind me that just having R&D funding for NonStop is testament alone to the true strengths of the product line – funding would have been cut off long ago if the NonStop business hadn’t returned to profitability. In talking with Fink I just couldn’t get over how, not once, was there anything other than enthusiastic support for all that is NonStop.
There’s much still to be written about NonStop and there will be many users surprised by the emergence of NonStop, the software play, but having covered NonStop for five years now I am pretty confident that I will still be behind the keyboard for at least another fifteen years.
“You get a variety of vectors that represent different thought processes and now we are recognizing the human issue; (within IT, there’s still) domain specialists - my server groups, that is, my Unix, Windows, etc. groups, my security, networking and storage groups, and so forth. How do I realign my human capital behind a Converged Infrastructure given how today we have many domain specialists? How is human capital to be trained?” Fink did suggest however, that “with clouds and virtualization then these (practices) are causing different groups to merge.”
Clouds? Virtualization? And even, Big Data? “Another complicated vector thrown into all of this (merging of domain specialists) and where NonStop plays – how we leverage NonStop expertise – is Big Data. People are talking about how they run big databases and with what they are doing is talking-up this idea of Massively Parallel Processing (MPP). And users will tell me ‘I probably need a columnar database (which NonStop isn’t) that’s on a MPP but then, wouldn’t it be good to bring in transactional data – a mixed columnar data store with transactional data’” Fink remarked before adding “but we have been running MPP engines that are some of the biggest the world has ever seen. There’s no better engine than NonStop!” While avoiding addressing any specific NonStop involved projects, and clearly rejecting any notion about NonStop being legacy, Fink did say that there were teams working on how best to leverage the intellectual property (IP) of NonStop with respect to Big Data.
As for Clouds, the projections were a little more cut-and-dried. “Most clouds today feature a lot of scale-out ‘pizza boxes’ and as long as you can partition the data, it all works,” Fink began before clarifying that there’s not any immediate future for NonStop inside the more popular cloud options being marketed. However, when it comes to NonStop, “being on the edge (of the cloud as) the resilient cloud presenter,” Fink then added, may prove to be an option for users already familiar with (and invested in) NonStop. But then pragmatically, with clouds that exist today we see just a further iteration in the long-lived cycle of technology trends with what we are see being just another phase as users continue to pursue less-expensive processing options.
Looking ahead to what the NonStop community can expect to see over the course of the next two to three years, Fink was even more pragmatic, saying “it’s all about Poulson. And looking at what it does for our users. It (provides) another performance leap and what we shouldn’t ignore is that it’s not just another chip upgrade but the first major architectural revision to Itanium since (Itanium 2) McKinley delivered in 2002.” And according to Intel, Poulson “will use a new microarchitecture, with a more advanced form of multi-threading that uses as many as four threads, to improve performance for single threaded and multi-threaded workloads.” Or, as Fink reiterated “it’s everything we see today, plus more (that we can use) to deliver even more for the NonStop user – it will be the high point for NonStop in the coming years.”
The image at the top of this post of me standing in front of an amusement park featuring wild rides perhaps isn’t altogether a fair reflection on NonStop. Many within the NonStop development organization remind me that just having R&D funding for NonStop is testament alone to the true strengths of the product line – funding would have been cut off long ago if the NonStop business hadn’t returned to profitability. In talking with Fink I just couldn’t get over how, not once, was there anything other than enthusiastic support for all that is NonStop.
There’s much still to be written about NonStop and there will be many users surprised by the emergence of NonStop, the software play, but having covered NonStop for five years now I am pretty confident that I will still be behind the keyboard for at least another fifteen years.