I am in full-on preparation for an upcoming webinar I will be delivering jointly, with IR's Jack van Meel. The presentation has been titled, House of Cards: NonStop X & Hybrid Computing and while it's something I am looking forward to, the topic has aroused interest over just how different the NonStop X is proving to be. This is not to say it's a case of bad news but rather, there is now much to be excited about in this latest implementation of the NonStop architecture. As we all are becoming familiar with its capabilities I feel that NonStop X is just a good deal for everyone in the NonStop community.
Jack and I really would welcome your participation and indeed, look forward to seeing you among the attendees. I will also be at the upcoming NonStop Technical Boot Camp (TBC) and will be only too happy to discuss the finer points should time detract for the level of in-depth discussion you would like to pursue.
Once you get past the basics of what's new with NonStop X there are some interesting changes worth considering should you be looking to introduce NonStop X systems into your data center. Of course, there's the successful transition to the Intel x86 architecture that has attracted considerable attention, even if it was all rather predictable following the announcement of Project Odyssey. It was four years ago, this month, that the announcement was made, "HP to Transform Server Market with Single Platform for Mission-critical Computing - Expanded HP Converged Infrastructure delivers industry-leading choice, investment protection."
According to then senior vice president and general manager, Business Critical Systems, HP, Martin Fink, "HP plans to transform the server landscape for mission-critical computing by using the flexibility of HP BladeSystem and bringing key HP technology innovations from Integrity and HP-UX to the x86 ecosystem. Unlike the competition, HP offers an open, integrated, single platform approach." Fink has moved up the organization in the four years since Project Odyssey was unveiled and is now the HPE EVP and CTO for the entire organization and his penchant for all things open and industry standard is matched with his ongoing, and indeed enthusiastic, support of NonStop. If you are unaware that Fink will be providing the keynote presentation at TBC, it should help foster greater interest in attending this event as I am confident Fink will not be holding any of his punches in front of such an audience.
However, NonStop X is far more than just the adoption of x86, or even the embracing of industry standard InifinBand (IB) as a logical follow-on to ServerNet, as it represents a new round of investment in the NonStop architecture. Many within the NonStop community had been expressing concerns about the future of NonStop given that roadmaps of the day stopped with the final chip spin for Itanium. It's as if that nasty proprietary shark has been chased away and it's OK to go back into the water where only open, industry-standard, solutions can be seen swimming with the current. All those decades of expertise developed in support of NonStop now has value and enterprises will be well served by the in-house expertise they can now fully leverage.
The introduction of IB does raise a couple of issues though pertinent to how the data center pursues connectivity and how it then applies monitoring solutions to the networks that develop. There's been a lot written about hybrid computers and networks and this will come to the fore when NonStop X systems are deployed at enterprises already running NonStop i systems (or even earlier systems) that wish to keep these older systems as an integral part of their operation. In particular, those NonStop systems with ServerNet will face the prospect of there not being an appropriate switch that can interconnect ServerNet with IB - Expand/IP over high speed Ethernet being the alternate - and should an existing system be paired with a 10Gb Ethernet NIC, then the subsequent pipe connecting these systems to NonStop X will be faster than what may already have been in place, but not IB speeds.
Furthermore, should WAN support still be necessary, then NonStop development has elected not to bring across major networking subsystems, including such important stacks as SNAX, preferring instead to rely on more modern WAN stacks from the likes of Infrasoft (uLinga) and even ACI (ICE) – HP's stated preference being uLinga. New interconnect and new network stacks may prove challenging but this could all be eased significantly with good monitoring solutions such as Prognosis. Apart from the more obvious reality of IR and Infrasoft both being based in Sydney Australia, there's also the effort IR has put into testing Prognosis on NonStop X at the HP Advanced Technical Center (ATC).
Tens and hundreds of hours of testing have been completed and if you would like to know more about the results, then make sure when you attend TBC, you check in to the California-60 meeting room at 5:00pm Tuesday for the presentation by van Meel, "Migrating from Integrity/Blade servers to NonStop X" for the inside scoop on all that was encountered migrating Prognosis to x86. Arrive at the meeting rooms an hour earlier go to the Gold–120 meeting room where Jamie Pearson will be presenting initiatives within Prognosis that will be "Bringing Analytics to real-time monitoring – the best of both worlds". And of course, there's a vendor track featuring IR that will be held at 10:30am in the California-60 meeting room.
The recent decision by HPE to exit public clouds and to focus on private clouds and indeed system hybrids that include legacy IT coupled with open systems and private clouds marks a distinct change in direction for HPE. According to Meg Whitman, in an interview with CNBC following the opening bell (which she rang), the private cloud market is much bigger than the public cloud market and besides, it's much better for HPE to consider cooperating with the likes of AWS than to play catch up. Furthermore, added Whitman, the vision of HPE will be on a hybrid cloud where some portions of the hybrid will be locked-down on premises but with some apps running on public clouds so no, HPE isn't abandoning clouds, HPE is simply going to double-down on where it can win in the marketplace and that is with support of hybrids and hybrid clouds.
There really aren't any surprises with this as the industry has been made well aware of this vision for some time. But what hybrids implies is mixes of systems and this is where application performance monitoring products capable of supporting a variety of platforms as a single business resource will benefit from products such as Prognosis. Increasingly, data center managers will want to see interactions between systems just as business managers want to check on SLAs even as they span multiple systems. To this end, Prognosis is really at the fore of providing such support and this will be covered in the upcoming webinar, where I will have even more to say about NonStop X and Hybrids.