Two major milestones were reached this week for fans of ARM-based server gear. First, HP and Applied Micro announced the 1st production ARM server this week with appropriate fanfare. Here’s an analysis by Paul Teich and Gina Longoria (another Calxeda Alumni) of Moor Insights and Strategy. Second, AMD showed off 2 OS’es (RedHat and SUSE), 2 JVMs (OpenJDK and Oracle), and Hadoop running for the 1st time on an ARM A57 based server at JavaOne. As Harish Jonnalagadda of BSN noted, “Adding Hadoop functionality to its software ecosystem is a natural move for AMD as its target clients will be looking to use its server clusters to process large data sets. The low-power nature of the CPUs make them ideal for processing large chunks of information and undertaking high I/O tasks.”
The HP event has been a long time coming, and heralds the beginning of a new era of choice for the datacenter looking to lower costs and power consumption. Since HP 1st announced Moonshot as a project with Calxeda in November, 2011, there has been a lot of press, and a ton of PowerPoints, but the market has had to wait for nearly 3 full years to see the vision finally begin to materialize for a production ARM server. Was it worth the wait? Honestly, this looks like a really good start. Assuming Gina and Paul did their homework well (and thats a pretty safe bet), it bodes well that HP can achieve a 35% TCO reduction in a 1st generation ARM product. Once 28nm ARM Server SOCs are available from Applied, AMD, and perhaps others, we should see power drop significantly (around 40% by my math), and the performance will improve as well.
So now, an enterprise class, production ARM server can finally be ordered, installed and deployed in a production environment. I congratulate Applied for getting to market first, and HP for having the vision and courage to bring this to market in a meaningful way. In addition to the initial support for Canonical Ubuntu; RedHat’s ARM project is widely expected to deliver a supported OS sometime next year, about the same time other (28nm) SOCs become available. While the power envelope of the 40nm xGene 1 limits the implementation to 1 SOC per cartridge, 28nm alternatives should improve that by at least a factor of 2X next year. Many, however, will lament that we still lack a low-cost development platform. Starting price of a Moonshot development box is well north of $50K. Hopefully this is something that HP or Dell or another player will address soon.
Disclaimer: (Sorry, but I need to say that …..) All content provided on this blog is for informational purposes only, is my personal opinion, and does not necessarily reflect the views of my employer. The owner of this blog makes no representations as to the accuracy or completeness of any information on this site or found by following any link on this site. The owner will not be liable for any errors or omissions in this information nor for the availability of this information. The owner will not be liable for any losses, injuries, or damages from the display or use of this information. These terms and conditions of use are subject to change at anytime and without notice.