ActiveWin.com: HP Superdome Technology Day in Cupertino, CA
HP decided to celebrate the 10 years of HP Superdome and market it through the HP Superdome Tech day with independent bloggers by inviting them to their Cupertino office in California on October 5th. Among those present were Shane Pitman from TechVirtuoso, David Adams from OSNews, David Douthitt from Unix Administratosphere, Andy McCaskey from SDRNews and Ben Rockwood of Cuddletech. The Tech Day was structured around the release of 3rd generation of HP Superdomes and to explain the features and benefits of the gigantic super-machine. There was a long list of HP panelists who were present at the event to help us get more information about the product. The day started with a short welcome from the VP of BCS marketing – Lorraine Bartlett and then an introduction to the HP Superdome range (Past, present and future) by Jeff Kyle who is the BCS Server Marketing Manager. For those who are uninitiated to the Superdome, here’s a quick link - http://h20341.www2.hp.com/integrity/w1/en/high-end/integrity-high-end-servers.html. The Superdome is HP’s reply to the mainframes; in particular the latest Superdome is the ammunition against the Z10 range of mainframes. It is the high end of HP Integrity servers which can run either the Itanium or the PA RISC architecture. Gigantic yet very simple machines, which I will explain in detail later.
Following Jeff was Brian Cox – the BCS Director for Software Marketing, who spoke to an interactive crowd regarding how to ‘unleash the capabilities of Superdome with HP-UX’. Now we do understand that it’s not a discussion that us Windows users will like to deep-dive in. However, it did bring out some good aspects of the Superdome’s capabilities. I asked for the roadmap on the Windows side of things and I was told that someone will get in touch with me and mail me a copy of the roadmap plan. When it reaches me, it reaches you.
After the talk by Brian about HP-UX and Superdome’s happy married life, we got ready for some real action and live demos. The reason I say those two things in the same breath is because once we walked in and saw the simulation (demoed by Richard Warham), it brought out some really good aspects of the HP Integrity platform. The war room scenarios that he generated showed a couple of great capacity planning examples. It was interesting to see the capacity being managed (read ‘increased and decreased’) on the fly as requirement/user load went up and down. Unfortunately, this dynamic allocation is possible only on HP-UX. So, I asked the question – what about Windows?? We know that HP Service Guard (which btw, is the HP’s internal clustering system and was making all this possible) does not work on Windows environment. So how do we get these features (if we can) on a Windows environment? The answer was surprisingly honest and quick – there’s no direct way of doing this on Windows. But there is a workaround – virtualization (Aha! The savior cometh). So apparently, HP-UX as a base OS, with a virtual windows servers on top of it will be able to do this – which translates to – if you don’t want to give us your money by buying HP-UX, you have to pay us the money and pay VMWare the money and then pay Microsoft the money before you can have what we give. And then the money question was brought up again – With the modularization of all HP product sales, how many modules of how many products we need to buy to get to this end product – and the answer was – it’s packaged. The cost is now in terms of a packaged solution. So well, the situation was getting better – from completely alienating the Windows users, they have started to at least glancing towards the 40% market that believes that Windows servers are good solutions for datacenter usage. After the demo, and a lunch break, we went back in for a walk-through of their lab where we saw the first generation Superdome running its third generation grandsons architecture and chassis. Some of the things that were still on paper were finally confirmed.
The superdome has an advantage over its competitors primarily because it is a very simple machine. For an untrained eye – it will look like a bunch of blade servers in a rack thrown in vertically. Cooling on top, power at the bottom, and extensions on the sides – Cleanly organized and everything visible. A bunch of plug and play hardware, if a cell (one server out of the range of blades that were thrown in) stops working, shut it down, pull the power, and pull it out, without unsettling the rest of the bunch. At the max, it has a capacity of 2 TB RAM (I think I remember someone saying that they are in process of making it 4 TB).
Overall, the HP Superdome seems to be a superior product to the competition which is the z10 Mainframe. One day full of a bundle of information, though we had a few fun moments when the Solaris advocate in the Sun Microsystems shirt tried to bash the Superdome in comparison to the Starfire E10k in the days when both had arrived to the market about 10 years ago. From HP’s side, it was a new tactic to reach out to the social media, and get themselves feedback from the rest of the world. For us, it was a good day with another new technology (even if it is an upgrade to the previous generation). And for all of you (our readers), it is an introduction to what’s to come in the war between HP and IBM in this business.