Intel is open-sourcing the code that enables resource pooling in infrastructure systems such as Ericsson Hyperscale Datacenter System 8000, the first commercial implementation of Intel Rack Scale Design. Intel® Rack Scale Design, formerly known as Intel Rack Scale Architecture, will be available on July 28, 2016.
The largest hemispherical building on Earth is the Ericsson Globe, in Stockholm. It’s a gigantic white ball with a 110 meter diameter. Looking at it from the outside, you’d have no idea that its primary use is as a stadium for ice hockey.
That’s the difference between architecture and design. The architects, Svante Berg and Lars Vretblad, asked the question, “Broadly speaking, what structure could I create that would facilitate the best use of its space?” And a designer, standing inside the giant ball, would later ask, “How can I fit a hockey rink inside this thing?”
Explore research by Mainstay on how hyperscale datacenters show over 50 percent savings in CAPEX and OPEX
The difference between architecture and design might also be understood in terms of Intel’s recent announcement that their platform, formerly known as Intel® Rack Scale Architecture (Intel® RSA) is now known as Intel® Rack Scale Design (Intel® RSD).
Intel® RSD makes your datacenter agile and lean
Intel® RSD is a paradigm for making datacenters operate, scale, and refresh more efficiently, by leveraging the best techniques in networking, virtualization, and resource pooling. Put another way, Intel® RSD is a way of simplifying the advanced technology that is used to accelerate the adoption of open, interoperable solutions of right-sized hyperscale datacenters.
Traditionally, datacenters were vast warehouses filled with racks of aggregated servers. Which is to say, every computer in the building had a pre-defined set of components—CPU, NIC cards, memory modules —and all those components were physically assembled into one hardware system. If a certain CPU was used for a certain process, and that process did not require all the associated memory modules, then that extra memory was wasted.
Using Intel® RSD, Intel empowers its partners to separate the Intel-made CPUs from the rest of the components—both physically and functionally. Any given process uses the resources it needs; no more, no fewer. When Intel comes out with their next generation of CPUs, those CPUs can be added to the resource pool without disruption. There is no need to throw out the components associated with the old CPUs; they too, will remain part of a resource pool.
How Ericsson is implementing Intel® RSD
Ericsson was the first partner to announce a product based on Intel® RSD (at the time, Intel® RSA): Ericsson Hyperscale Datacenter System 8000, which was introduced at the 2015 Mobile World Congress (MWC). The form factor of Ericsson Hyperscale Datacenter System 8000 makes component pooling easy, its optical backplane enables light-speed data links, and its data network's folded Clos spine/leaf topology eliminates bottlenecks and reduces latency.
With the Ericsson Hyperscale Datacenter System 8000, Ericsson leads the industry in shifting to a paradigm in which the physical proximity of components is less important than their functional relationship.
Case in point. One day in May of 2009, a small red cottage, in the tradition of the aesthetic of the Swedish countryside, appeared on top of the Ericsson Globe. It was an artistic installation by artist Mikael Genberg. Actually, most of Genberg’s art consists of putting red cottages in unexpected places. In doing this, he urges the observer to contemplate the juxtaposition of conceptually similar, yet functionally disaggregated, physical structures.
What Intel open-sourced and where to find it
This week, to encourage a similar type of contemplation (but in the realm of agile infrastructure design), Intel is making the code associated with Intel® RSD available as open source. The code to be released is associated with the following components: the Pod Manager (PODM), the Pooled System Management Engine (PSME), the Remote Management Module (RMM), and Vtune (VTS). It’s ready to productize and incorporate into orchestration solutions.
With this open source release, Intel provides a clear path to address datacenter management fragmentation, drive down Total Cost of Ownership (TCO) via resource pooling, and improve the management of operating expenditures (opex) via enriched telemetry.
Download the code from GitHub now to begin exploring how you will innovate with Intel® RSD.
Read about how Ericsson has incorporated Intel® RSD into Ericsson Hyperscale Datacenter System 8000.
Read about it on the Intel blog.