Redstone on the road, in the lab, and playing at a cloud near you: HP Moonshot Update

[#1]

Guest blog written by Glenn Keels, Director, WW Hyperscale Product Marketing, HP

Since our last update, the Moonshot team has been especially busy.  Things getting really interesting and fun.   This is especially true for the collaboration we’re seeing with our Pathfinder Members.  Here’s the low down on what we've been up to lately!

Within days of that announcement Redstone was seen being powered by RedHat’s Jon Master’s bicycle.  Some might say; “cute marketing stunt”  – I say it shows that our joint vision is becoming a reality.  Jon noted that he actually could power a server with his bike before.  Now, he can power more than 200 in a single Redstone system.   Check Jon’s blog out here.

The next couple of weeks were breakthrough.  By July 1, Our HP and Calxeda engineering teams had stood up enough Redstone systems up in the Discovery Lab to welcome our first outside visitors for in depth testing and development.  The company honored to be our first Discovery Lab engagement was Canonical.  We all learned a tremendous amount.  I personally am excited about the possibilities that Juju and MAAS hold for the future. Perhaps Mark and Chris from Canonical would like to comment and share their excitement?  It didn’t stop there though.  Redhat was hot on their heels in the lab next.  It’s simply great to finally see the work of over six months start paying off with real engineers, real Moonshot servers, and real progress.

Just three weeks ago, Calxeda came up with the idea that while we’re standing up more Redstone systems in the Discovery Lab, we could make some available in the cloud for those waiting.  Great Idea!!! Together with Canonical and our friends at OpenStack and CoreNAP, we stood up the first TryStack zone for ARM in short order. And in the first week, had some really good initial usage numbers.  We think this is really working to spread the excitement of OpenStack and ARM Servers... and it’s FREE!  We’ll try to stand up even more trays for expanded capacity in the coming weeks!

Redstone is becoming a reality and our Pathfinder members are making it happen.  We've received overwhelming interest in Moonshot and Discovery Lab and recognize many customers are patiently waiting for systems to be available through the Discovery Lab.  To all of you, we appreciate your patience.  We’re working hand in hand with Calxeda to stand up enough systems for the significant demand as well as to prioritize shipments to the handful of customers requesting them on site.  We’re almost there and will update you here and via email.  In the meantime, if you just can’t wait to get your hands on a system, I really encourage you to join the ARM TryStack Zone through Facebook and take some for a free spin.  And don’t forget to let us know what you think!!

On another note, we've also made a lot of progress on our Pathfinder Partner Program. I was in the airport last week with Tim Wesselman from our CTO’s office.  Tim runs Partner Strategy and the Pathfinder Program.  One of the things we realized was that many people don’t realize how powerful the collaboration is within the Pathfinder Program already.  And we want more.  Tim, would you mind commenting on that as well as the status of our membership?

Best regards,

Glenn

For more information, read more about HP Project Moonshot, or join the HP-CAST LinkedIn community.

[#2]

I can see why this excites you Glenn, but what about it should be exciting to the rest of the world? How will it shake up how computing is used to generate value for businesses??

[#3]

but what about it should be exciting to the rest of the world?

Power is very expensive in a Datacenter. Here's why it's different than your usual home power supply :

( 1 ) Multiple power-provider : you need to ensure continuous supply of power even if one of the power company went down. This is not easy to setup, and it will cost extra.

( 2 )  Special equipment for point 1 above : to ensure smooth transition to other power supply in case of failure in the primary power line.

( 3 ) UPS : Uninterruptible Power Supply : this devices will ensure a clean, spike-free power supply to the servers. And it will provide power from its batteries in case of total power failure, while waiting for the generator to kick in.

And, if you thought the UPS for PC / home are expensive, wait until you see the prices for Datacenter-grade UPS.

( 4 ) Generators : in case of total power failure, the generator will kick in and provide the juice in the mean time.

The bigger the generator, the more expensive it is. Simple & straightforward.

( 5 ) Air Conditioning : one thing that most people misses is the power requirement for the A/C. Without the A/C, most servers will overheat in just several minutes !

Not only the UPS & Generator must be of enough capacity to power the servers - they also must have enough capacity to power the A/C in case of total power failure.

So as you can see, if you can bring down the power requirements & heat output, then it will translate directly to some serious cost savings. 

It is very worth noting that these ARM-based servers produces much less heat than Intel servers. 

You can see in the pictures, they don't even have a heatsink tacked on. I'd venture a guess that they may run just fine even in room temperature == no / minimum cooling == major cost savings. 

Less / no reliance (to cooling system) == increased reliability.