With seven posts to
the LinkedIn group, Real Time View, written while attending 2013 HP Discover, a
week later I have had time to reflect on all I saw. And it really is game-on
for NonStop!
Readers may recall that our trip last year to Las Vegas was via Rosamond, California, where we kicked off the week with a weekend laying down laps at Big Willow – much of this covered in the post of May 31, 2012, Java rocks! But no, this time it would be completely different.
As 2013 HP Discover wound down, we headed for our company Command Center , our transportation hub, this time just as it was last year, and headed north to San Francisco and to Sonoma, California. We spent the weekend laying down laps on the famous track at Sear’s Point. However, this time I crashed, backing the car into one of the track’s infamous concrete barriers – more to follow, of course, in postings to other blogs. The driver is quite OK, apart from being slightly embarrassed, but the red ‘Vette is in the shop having its rear end corrected.
On the other hand, I was invited once again to participate in HP Discover as part of HP’s supported blogging community, and HP put on quite the show and one that proved well worth the effort to attend. If last year’s event could best be described as being steady-as-you-go with HP CEO, Meg Whitman, only being at the helm for a few months, this time it was a more assertive Whitman who appeared to be giving the command to the rest of HP to lock and load! It is time, apparently, to lead with the front foot and to come out swinging.
There are risks involved in every pursuit. Nothing comes without a price being paid – to be truly good at anything requires considerable effort as well as patience and perseverance. To stand out, to innovate, to change the game – that requires revolutionary thinking, and on several occasions we saw HP demonstrating a new-found swagger that was good to see.
I have been driving cars on racetracks now for six years and am gradually coming to terms with what it takes to drive well. But the weekend’s crash only reinforced how much further I have to go and just what risks are involved. I cannot stress strongly enough that the association I belong too goes to great pains to ensure everyone’s safety while on the track, however, as experience increases so do speeds and with speed comes a need for precision and it only takes one mishap to produce devastating results.
In business, managing risk is a fundamental cornerstone for senior executives – bet-the-business situations are rarely undertaken. Books have been written over the years whenever such a situation took place – Boeing bet-the-business on the Boeing 707 jet and came out victorious, but so many other companies simply crashed and burned. Remember Osborne with it’s prematurely announced the Osborne Executive? However, this time we saw firsthand just how powerful an initiative Converged Infrastructure has become and how big a game-changer Moonshot would be.
For the NonStop community much of what these major initiatives represent seems far removed from daily concerns – indeed, some of the goals of these initiatives look to be playing catch-up to what is already available today with NonStop. Commodity hardware? Check! Standard languages and tools? Check! SQL? Check! Simply building block architecture with nothing shared? Check!
Hold on just a tad – simple building blocks? Shared nothing? The more I looked into a drawer full of Moonshot cartridges the more I saw the future, and with it, a possible glimpse of what very well might keep NonStop relevant for decades to come. The Advanced RISC Machine (ARM) processor and it’s closely-related peers are taking us back to a much simpler approach and I like that a lot – yes, CIOs polled of late are all looking to simplify technology.
“The ARM processors are all single core,” was how one HP colleague expressed it. “The good news about single core is that it really devalues multithreaded applications – no need to complicate the core if you only have a single core. If programming models were to go back toward single or minimal thread applications, the NonStop starts looking very good again and applications should port much more easily and have immediate scale through Pathway server class cloning.
Of course, the new Intel ATOM chip will likely see additional cores added – but the key here is that the foundation for all these chips is the mobile phone / tablet industry and with cores come heat. The mobile phone / tablet users despise heat and would reject any future product if it was simply “too hot to handle”. So perhaps the rush to stack cores as high as possible has run its course. And if so, my HP colleague may very well be right.
For now, the future for NonStop is very secure and the roadmaps call for further porting to Itanium Poulson and Kittson chips. There’s very little risk associated with this roadmap – more steady as she goes philosophy, and a continuation of the program to provide increasingly better price / performance metrics. But looking further down the road, there’s obviously going to be increased risk.
Will there be a new iteration of the x86 chip that combines the robustness of Itanium with the wealth of applications that exist today for current generation of x86 chips? Or will the arrival of Moonshot, with ARM / ATOM create a very large fork in the road where coins will be tossed, and chickens gutted? Will it be another case of lock and load? The very real prospect of IBM selling off its server business almost sounds as if they knew something really new was heading their way and weren’t prepared to double down to compete.
Earlier this year, in an interview with e-publication, ZDNET.Net, Whitman told the reporter of how “there are a number of new programs and disruptive innovations that should help us along … we will be bringing the latest innovation from HP Labs and our Enterprise group to market, the first commercialized product from our Project Moonshot. We expect this to truly revolutionize the economics of the data center with an entirely new category of server that consumes up to 89% less energy, 94% less space and 63% less costs than our traditional x86 server environment.”
HP and disruptive innovation? HP and revolutionizing the economics? This is definitely the new HP, and one apparently quite prepared to take risks. For all who participated at last week’s event, it was impossible to miss just how proud all involved at HP were with what they have achieved in just twelve months – yes, much of what was revealed had been on the drawing boards and in the labs for much longer, but little had leaked out to prepare us for what eventuated.
It will be only a couple of days for the red ‘Vette to be ready for the track again and we have a big weekend planned for mid-July. For HP’s competition, it will not be that simple. I just cannot see where competing technology will come from – yes, Project Moonshot is in its infancy and is targeting simple web and proxy server “applications”, but over time it holds the promise of so much more with every likelihood of handling the many different workloads typical of today’s enterprises. And NonStop will not be denied in the years to come!
Comments
However, while smartphones and tablets are highly successful and clearly have their strong domains, they cannot completely replace the PC - they are optimized for consuming content, but are less suitable for creating and editing content of more complex nature.
So, what type of device will take that role in the future ? Maybe we will see dumb terminals again, of course these would be browser terminals providing all the rich display and pointing functionality we know from today's PC's. However they would be a lot easier and cheaper to manage, more reliable and much more secure. And they would be optimized for cloud computing. Microsoft have already made their Office functionality available in the cloud ...
Would those cloud computing terminals still be driven by X86 CPU's ? Not necessarily, there is no reason why those cheaper ARM parts we know from smartphones and tablets could not handle a keyboard and a mouse too.
The X86 architecture has come a long way - it was introduced back in 1978 with the Intel 8086 microprocessor. It won't disappear over night, but somehow seems to be running out of gas. Decades of bolting on further functionality resulted in high complexity, improvements in CPU power don't come so much from clocking faster but from adding more cores - thus further increasing complexity.
Chip architectures also have their lifecycles, the Alpha - unbeatable for its speed back in the 90's - has reached end of life much quicker than X86. But it also seems quite clear that X86 won't dominate forever.
Today's prevailing system environments with their many components added in a mix-and-match fashion are just too complex and hence, not reliable enough for the Internet age. In contrast, NonStop is a very simple and robust system providing all the reliability, scalability, manageability and security needed in the era of the Internet and cloud computing. And NonStop has a proven track record of changing chip architectures when needed.
There is no need whatsoever for such a change now, as it is decided that NonStop will run on the next two generations of Itanium. But there is no question that NonStop can and will move when some day a even better suited chip architecture emerges.
Will ARM some day develop into such a better chip architecture suitable for critical servers ? Much too early to say at this time - but while ARM might be ready sometime in the future, NonStop is ready now.