The timelessness of NonStop on display as Intel's x86 is embraced - yes, the plan to support of x86 architectures as part of the NonStop product family has been unveiled at this year's NonStop Advanced Technical Boot Camp.
While I was visiting Australia, I was kept informed of a
young up-and-coming racecar driver. While only in his early teens, he was
featured on American television as one of three youngsters worth watching as
likely future stars. Colton Herta, the son of Indy Racing League (IRL) team
co-owner, Bryan Herta, this year raced in two programs for open wheel racecars
(smaller versions of F1 cars). In doing so, he managed to win the major series
(in an F1600) while narrowly missing out, as we heard while travelling, on
winning the second series (Skip Barber formula) by a margin of two points even
though conflicts with the F1600 series meant he missed many Skip Barber events.
For 13 year old Colton, campaigning against much older participants with years more experience in open wheelers, winning has become a way of life so much so that his trophy room now only houses those trophies awarded for first place. Being good is no longer the point for Colton and just being good enough on the day, never a consideration. It’s also a fact that fellow racers are going to have to get used to as the picture above depicts. (The other racecar driver featured on American television alongside Colton? Turned out to be Matthew Brabham, grandson of Australia’s own Sir Jack Brabham.)
At a time when it’s becoming apparent to many IT professionals that perhaps, with budgets trimmed as tightly as they have been for the past four or five years – ever since the Global Financial Crises (GFC) –good has become good enough, all involved with NonStop have become perplexed over the almost daily reports of major system crashes. From government data centers to stock exchanges to online retail service providers, we all watch as major failures are simply dismissed as being nothing more than technical glitches and that an acceptable response from those in charge is that everything is now OK as they powered-off, powered on, and reset the application.
The humble PC has indeed revolutionized IT but it’s not without its downside not the least being it’s influence on IT managers and executives who rationalize that every need in the private and public sector can be addressed with essentially a PC-based server. No; good by itself, is simply not good enough! Unfortunately, the discussions about doing a better job for IT usually stop there, as few are willing to set aside the budget required to properly address these availability shortcomings and for further transformation of IT to take place, bold actions must be taken by the major vendors.
It was five years ago, on March 26, 2008, that I posted A question of balance! to this blog. In that post, I quoted Randall Becker of Nexbridge Inc as stating, at the time, how with NonStop, “HP has a real diamond here, if they could only see that the cut glass, through which they’re looking, isn’t good enough!” Randall then asks “but then, we all struggle with the question of what reliability is good enough? Where do we set the bar?”
A short time ago, I posted NonStop: Good, but now better! where I concluded that when it comes to the merits of NonStop few discount the value the presence of NonStop can provide. I then made the observation that, as we continue our march to even greater heterogeneity in the data center (with hybrid computing becoming common place), there will be lot of good systems inside the data center but the better solution for mission critical transaction processing will continue to come from NonStop!
As more information became available for HP’s project Odyssey, HP talked openly about extending the full mission-critical experience for x86 and all within a common modular infrastructure. At the time, there were presentations about how the knowledge within the NonStop development team would be exploited by developers working on other OS’s. However, at no time was there any discussion about NonStop bringing its availability proposition to x86. Quite the contrary, NonStop would remain positioned at the top of the pyramid depicting HP’s product lines – it’s unique attributes reserved for those in IT well-acquainted with NonStop’s true value proposition. Yet many within the NonStop community wondered whether there would be a time when NonStop did add support of such popular chip architectures as the x86.
It was in 2012, following that year’s HP Discover event, in the post Sailing, tacking and avoiding conflicts! I referenced Intel. For me, I posted at that time, it has always been the “Intel Architecture”. It was less about the chip products per se and in I referenced a conversation I had with Intel Corporation’s General Manager, Enterprise Software Strategy (and former head of NonStop Enterprise Division), Pauline Nist, who reminded me that whatever transpires with future chip products, the Intel Architecture will prevail. In telling me this, Pauline was reiterating clearly that NonStop customers would be protected and upward compatibility would be assured. Should we anticipate NonStop supporting Itanium and Xeon going forward? Two distinct marketplaces with two separate product lines?
The question over good becoming good enough has bothered many within the NonStop community but this week, at the NonStop Advanced Technical Boot Camp, HP demonstrated that indeed they will be taking the bold steps. The bar hasn’t been reset – being good isn’t good enough – but rather, rising up to the height of the bar will now be more attainable as HP NonStop embraces x86 and the economic model that comes from making this move. That’s right – a complementary product line exploiting for the first time x86 architecture with InfiniBand. Welcome, too, to the new Lxx operating system.
Imagine a future, we have just been told, where you could provide customers and business partners alike, 100% NonStop fault-tolerant environment on the Intel Xeon platform! “Our NonStop customers truly make it matter,” said HP CEO Meg Whitman in a video adding how pleased she was to extend NonStop to include x86. Demonstrating once again for the NonStop community that indeed, NonStop “is a timeless architecture,” was how HP VP and GM of Integrity Servers Randy Meyer summed it up. Key objective of this program (with deep roots into both Project Odyssey as well as Converged Infrastructure), according to one NonStop manager, was “maintaining compatibility at the application level – just recompile and go”!
The move to InfiniBand has been predicted for some time. In my post of February 13, 201, Three years on, and three more wishes! I quoted HP CTO Martin Fink who had proposed that, “with every new microprocessor that becomes available to us, we continue to evaluate ServerNet and its impact on overall system performance. Certainly at some future date, we'll move to a standard interconnect – probably Infiniband as a possible alternative.” Now that Fink’s observation has come to fruition, not as a replacement to ServerNet but as a complementary fabric solution better optimized for use with x86 architectures.
In the coming weeks as we hear more about these new technologies and products are I will be following up with further posts to this blog. Young racer Colton Herta may be comfortable with winning and not one to accept good being simply good enough, but I suspect it will take some time before the rest of IT wholeheartedly embraces such a concept. With HP delivering, as we have just heard, “HP NonStop on the industry’s leading platform and interconnect, (it) means your transactions will execute with even faster performance and lower latency”.
Furthermore, “The combination of HP’s proven leadership in mission critical x86 technology and in HP NonStop technology will offer a compelling business choice,” HP pitched in its presentation today further highlighting how big a transformation of NonStop has occurred. With HP saying this, the many naysayers doubting the future of NonStop will find little to talk about in the months ahead. The bar didn’t get moved, and there was no temporary fix proposed for a select few (we didn’t get a box to stand on, for instance); this is the real deal. Believe!
For 13 year old Colton, campaigning against much older participants with years more experience in open wheelers, winning has become a way of life so much so that his trophy room now only houses those trophies awarded for first place. Being good is no longer the point for Colton and just being good enough on the day, never a consideration. It’s also a fact that fellow racers are going to have to get used to as the picture above depicts. (The other racecar driver featured on American television alongside Colton? Turned out to be Matthew Brabham, grandson of Australia’s own Sir Jack Brabham.)
At a time when it’s becoming apparent to many IT professionals that perhaps, with budgets trimmed as tightly as they have been for the past four or five years – ever since the Global Financial Crises (GFC) –good has become good enough, all involved with NonStop have become perplexed over the almost daily reports of major system crashes. From government data centers to stock exchanges to online retail service providers, we all watch as major failures are simply dismissed as being nothing more than technical glitches and that an acceptable response from those in charge is that everything is now OK as they powered-off, powered on, and reset the application.
The humble PC has indeed revolutionized IT but it’s not without its downside not the least being it’s influence on IT managers and executives who rationalize that every need in the private and public sector can be addressed with essentially a PC-based server. No; good by itself, is simply not good enough! Unfortunately, the discussions about doing a better job for IT usually stop there, as few are willing to set aside the budget required to properly address these availability shortcomings and for further transformation of IT to take place, bold actions must be taken by the major vendors.
It was five years ago, on March 26, 2008, that I posted A question of balance! to this blog. In that post, I quoted Randall Becker of Nexbridge Inc as stating, at the time, how with NonStop, “HP has a real diamond here, if they could only see that the cut glass, through which they’re looking, isn’t good enough!” Randall then asks “but then, we all struggle with the question of what reliability is good enough? Where do we set the bar?”
A short time ago, I posted NonStop: Good, but now better! where I concluded that when it comes to the merits of NonStop few discount the value the presence of NonStop can provide. I then made the observation that, as we continue our march to even greater heterogeneity in the data center (with hybrid computing becoming common place), there will be lot of good systems inside the data center but the better solution for mission critical transaction processing will continue to come from NonStop!
As more information became available for HP’s project Odyssey, HP talked openly about extending the full mission-critical experience for x86 and all within a common modular infrastructure. At the time, there were presentations about how the knowledge within the NonStop development team would be exploited by developers working on other OS’s. However, at no time was there any discussion about NonStop bringing its availability proposition to x86. Quite the contrary, NonStop would remain positioned at the top of the pyramid depicting HP’s product lines – it’s unique attributes reserved for those in IT well-acquainted with NonStop’s true value proposition. Yet many within the NonStop community wondered whether there would be a time when NonStop did add support of such popular chip architectures as the x86.
It was in 2012, following that year’s HP Discover event, in the post Sailing, tacking and avoiding conflicts! I referenced Intel. For me, I posted at that time, it has always been the “Intel Architecture”. It was less about the chip products per se and in I referenced a conversation I had with Intel Corporation’s General Manager, Enterprise Software Strategy (and former head of NonStop Enterprise Division), Pauline Nist, who reminded me that whatever transpires with future chip products, the Intel Architecture will prevail. In telling me this, Pauline was reiterating clearly that NonStop customers would be protected and upward compatibility would be assured. Should we anticipate NonStop supporting Itanium and Xeon going forward? Two distinct marketplaces with two separate product lines?
The question over good becoming good enough has bothered many within the NonStop community but this week, at the NonStop Advanced Technical Boot Camp, HP demonstrated that indeed they will be taking the bold steps. The bar hasn’t been reset – being good isn’t good enough – but rather, rising up to the height of the bar will now be more attainable as HP NonStop embraces x86 and the economic model that comes from making this move. That’s right – a complementary product line exploiting for the first time x86 architecture with InfiniBand. Welcome, too, to the new Lxx operating system.
Imagine a future, we have just been told, where you could provide customers and business partners alike, 100% NonStop fault-tolerant environment on the Intel Xeon platform! “Our NonStop customers truly make it matter,” said HP CEO Meg Whitman in a video adding how pleased she was to extend NonStop to include x86. Demonstrating once again for the NonStop community that indeed, NonStop “is a timeless architecture,” was how HP VP and GM of Integrity Servers Randy Meyer summed it up. Key objective of this program (with deep roots into both Project Odyssey as well as Converged Infrastructure), according to one NonStop manager, was “maintaining compatibility at the application level – just recompile and go”!
The move to InfiniBand has been predicted for some time. In my post of February 13, 201, Three years on, and three more wishes! I quoted HP CTO Martin Fink who had proposed that, “with every new microprocessor that becomes available to us, we continue to evaluate ServerNet and its impact on overall system performance. Certainly at some future date, we'll move to a standard interconnect – probably Infiniband as a possible alternative.” Now that Fink’s observation has come to fruition, not as a replacement to ServerNet but as a complementary fabric solution better optimized for use with x86 architectures.
In the coming weeks as we hear more about these new technologies and products are I will be following up with further posts to this blog. Young racer Colton Herta may be comfortable with winning and not one to accept good being simply good enough, but I suspect it will take some time before the rest of IT wholeheartedly embraces such a concept. With HP delivering, as we have just heard, “HP NonStop on the industry’s leading platform and interconnect, (it) means your transactions will execute with even faster performance and lower latency”.
Furthermore, “The combination of HP’s proven leadership in mission critical x86 technology and in HP NonStop technology will offer a compelling business choice,” HP pitched in its presentation today further highlighting how big a transformation of NonStop has occurred. With HP saying this, the many naysayers doubting the future of NonStop will find little to talk about in the months ahead. The bar didn’t get moved, and there was no temporary fix proposed for a select few (we didn’t get a box to stand on, for instance); this is the real deal. Believe!
Comments
If little-endian, it won't be compile and go, due to at least unstructured files that have big-endian data in them, and, to some extent, programs that have big-endian assumptions built in (sometimes unrecognized).
If a new big-endian mode on Xeon, that would be a fairly big block for quick ports of x86 software to the platform -- no less than there is now, but don't get excited about easy availability of all the x86 software in the world.
There's no new big-endian mode on Xeon. AS for your observations re data - well, I think that they have addressed that from talks with other vendors already briefed by HP. Hopefully someone will step in here although I seem to recall some responses to discussions on LinkedIn have covered this ...