HP is aggressively playing matchmaker for hardware vendors, software developers, and a broad range of customers as it seeks to establish its "make of it what you will" Moonshot hyperscale server architecture. "We're turning the development model upside-down, and we're starting with the solution," John Gromala, the director of …
Solution looking for a problem?
"We're turning the development model upside-down, and we're starting with the solution,"
That generally doesn't work out well.
Re: Solution looking for a problem?
".....That generally doesn't work out well." It doesn't work well if you try and bend the market to fit your solution (as Sun tried with CMT), but in this case hp seem to have taken a leaf out of Balmer's play book - "developers, developers, developers" - and extended it to "customers, customers, customers." The whole market is angling towards ultra-dense servers, it's just a matter of how us customers use them, so hp making their offering as flexible as possible makes sense. It's the methodology they used to take ownership of the blades market.
If you only pitch at the Fortune 2000, hardly a hotbed of innovation.
Re: justincormack Re: Especially
"If you only pitch at the Fortune 2000, hardly a hotbed of innovation." I don't think it's about being hotbeds of innovation. If anything, the financial houses want to be financial houses and would quite like to pay someone else to look after the IT, only regulation and competition drive them to be heavy investors in inhouse HPC-type computing. I know several maths grads that work on economic and market modeling in the City, they tell me they use much more complex HPC gear than was in use for research at their unis.
Having said that, I bet the first place hp took these was too all the hosting cloud companies.
Needs to be marketed differently
HP actually pitched these to us a while back -- we're a bog standard IT shop and my boss thought this was the next version of BladeSystem.
The truth is that this isn't your typical x86 server or blade platform. Customers who would use something like this probably have an idea of exactly what they need, and have a problem that needs a very customized system to solve. The first rollout of this was pitched directly to the Googles and Facebooks of the world, or more correctly, companies that thought they were a Facebook and Google and needed to get on the big data bandwagon. It looks like an attempt to stave off the white-box hardware solutions that the big social media and cloud startups have begun using in their data centers. It's targeted somewhere between the x86 blade customers (density) and midrange/Itanium customers (highly custom/complex systems) and it'll be interesting to see HP drum out some case studies to give customers ideas about what it would be good for.
Starting with the solution and then look for a problem - that really got SUN off the ground ;-)
That said, I'm really a bit puzzled as to what one can use these servers for.
The CPU doesn't look very fast, internal HD space is limited - what would one use this thing for?
Maybe an array for caching proxies with NGINX and Varnish?
Low-power, low-latency applications would be a prime application I would think. Call routing anyone?
What if Android takes off and you want android desktops but half the organisation has ipads? You run up remote desktops on headless, googleless android.
Sure, a single cartridge isn't impressive.
But when you tie a dozen of them together...
Its hardly needs any research - seems so easy to achieve:
Just needs some really crap plumber and an equally crap soundtrack for a guaranteed moneyshot.
Horse-power to chicken-power
Moonshot takes computing from horse-power to chicken-power. This is the exact opposite of the virtualization and consolidation that the industry has been embracing for the past decade. A decade ago we had a bunch of small and inefficient systems. Why buy a the custom HP cartridge and all the infrastructure to network them together? The alternative is to plug in 45 cell phones or tablets, throw them into a draw, and wireless network them together. This is like welding 45 bicycles together instead of building a car. When you weld 45 bicyles together, sure it uses less gas, and it's much lighter, and there's no complicated electronics, but you end up with something much less capable. A mini-grid of moonshot will be less flexible than one larger virtualized system.
If android is the answer, open source is the solution.
Re: AC Re: Horse-power to chicken-power
"Moonshot takes computing from horse-power to chicken-power. This is the exact opposite of the virtualization and consolidation that the industry has been embracing for the past decade....." Very true, hp need to be careful that their salesgrunts don't try pushing it as a solution to all ills, otherwise it risks being another Sun CMT. It's a niche product, maybe several niches, but not one design to rule them all.
- NASA boffin: RIDDLE of odd BULGE FOUND on MOON is SOLVED
- SOULLESS machine-intelligence ROBOT cars to hit Blighty in 2015
- BuzzGasm! Thirteen Astonishing True Facts You Never Knew About SCREWS
- Worstall on Wednesday YES, iPhones ARE getting slower with each new release of iOS
- Microsoft's Euro cloud darkens: Redmond must let feds into foreign servers