Today the SiliconAngle team is at HP, covering the announcement of Moonshot, an ARM based low energy server technology. Among the guests on our live show theCube was Forrester analyst (and former director of blade system strategy at HP) Richard Fichera. SiliconAngle’s John Furrier and Wikibon’s Dave Vellente talked with him about the announcement and what it means for the enterprise.
“We’ve basically been in an environment where it was assumed that you had an x86 instruction set and a processor from AMD or Intel, that’s basically the recipe that people have been working with,” Fichera said. He thinks the announcement of the use of ARM in HP servers is a big deal, though it won’t immediately displace the x86 market.
When asked why someone wouldn’t use low-power processors instead of an x86 architecture, Fichera said right now the workload characteristics are unknown, but over time once the workloads have been established there will be greater pressure to use these chips because it’s ultra-efficient.
As to specific usecases, Fichera said it’s unlikely to be used for CPU intensive activities like you derivatives calculation engines or structural analysis. But it would work well for things like streaming media, static Web pages or even lightweight dynamic pages. He said it would likely work for some mid-tier applications. For example, some Java workloads work well and some don’t. He says Hadoop will work well. Fichera says the software partnerships will help HP find out more about what actually works well on this platform, and it will mean that HP will be the first to know.
Fichera also notes that HP’s Redstone system is redefining what the system engineering stack looks like.
Advancements in server architecture and networking infrastructure have advanced relatively slowly compared to the rapid changes in software, the rise of cloud services and the explosion of data. The cloud is putting new pressures on data centers and networks, and it’s about time that we had a leap forward in infrastructure. Moonshot, along with technologies like OpenFlow, OpenCompute and solid state storage, could be the sort of progress we need.
Photo by s58y