Growing Faster than Moore’s Law: HP’s Project Moonshot

[#1]

Guest blog written by Gary Thome, Vice President, ISS Strategy  

By the end of 2013, there are forecasted to be 5 - 8 billion sensors attached to the web, with nearly every sensor attached in some way to a person wanting to collaborate and share information with others via text, tweets, attached email files and videos.  While this is an incredible number, it’s an unfortunate fact that existing computer architectures were not designed with this level of scale in mind.

In the past, software developers and IT professionals could find comfort in the predictable doubling of performance every 18 months (as is stated in Moore’s Law), which insured that future performance needs could be met.  However, new hyperscale web applications find themselves often growing much faster than the 18 month timeframe, and thus requiring more and more computing power. The challenge is that Moore’s Law is no longer good enough.  What is needed is a new computer architecture that can keep up with the performance and capacity growth of hyperscale applications, while at the same time not become overwhelmed by growth in infrastructure and power.

Moonshot Scales

This is where HP comes in.  On November 1, 2011, HP announced Project Moonshot.  This is a multi-phase, multi-year approach that is blazing a new trail for bringing extreme low- energy server technology to the data center.  Traditional computer architectures focus solely on ways to scale performance, yet what is needed is how to scale performance efficiently.  Traditional shared-everything and shared-nothing computer architectures both miss the mark.  In contrast, Moonshot is targeted at sharing almost everything – at scale points never previously conceived.  Of course, Moonshot is not for everyone.  In fact, we believe it will only be a fit for a relatively small number of applications, but these applications are the ones that run at hyperscale. 

As an ISS employee, I’m really excited about this project.  Changes of this magnitude don’t happen very often and this is a new paradigm in computing I believe will prove to be a critical component of the Internet of the future.  Pulling this off is no easy feat, but with our long history of computing innovation and strength at massive scale today, I believe we are best positioned to do this. 

Like many great innovations and explorations, including man’s mission in space, Project Moonshot requires collaboration across an entire ecosystem of industry leading partners, customers, and industry experts which will take some time to see full fruition, but we at HP believe the effort is worth it.

For more information, check out HP’s Project Moonshot for extreme low-energy server technology or other hyperscale information, or follow us on Twitter at @HPHyperscale.