Infiniband Addressing New Markets

Infiniband networking, typically used in high end supercomputers, including nearly half of the Top500 fastest supercomputers, continues to address new markets outside of the traditional supercomputing space.

Just today, NYSE Technologies announced availability of a significant performance upgrade of its middleware platform, Data Fabric, and demonstrated a message rate of over a million 200 byte messages per second over QDR (40Gb/sec) Infiniband. And as Data Center Knowledge reported earlier in the week, Microsoft’s Bing Maps site is now running on an Infiniband network.

It is easy to argue that 10Gb and new 40Gb ethernet technologies have broader market reach than Infiniband, and in fact the Mellanox CX2 cards used in the NYSE Technologies benchmark support both Infiniband and 10Gb ethernet, but Infiniband still has a clear performance advantage today when low latency is a key requirement. Meanwhile, Mellanox and Infiniband vendor Qlogic aren’t standing still. Mellanox is already selling CX3 kit supporting 40Gb Ethernet and FDR (56Gb/s) Infiniband, albeit to take full advantage of FDR you need to wait for new next gen PCIe-Gen3 compatible CPUs and servers to become available.

HP was an early adopter of Infinband technology and in fact designed Infiniband onto the motherboard of both our ProLiant SL390s G7 server and our ProLiant BL2x220c G7 server blade. In these designs, we used the Mellanox CX2 chipset and thus both products support 40Gb QDR Infiniband or 10Gb Ethernet without expensive add-on cards.

So no matter if your interests are in building one of the fastest supercomputers in the world, geolocation services, or simply trading stocks blindingly fast, HP ProLiant SL390s and BL2x220c servers are ready for the task. In fact, the world’s fifth fastest supercomputer, Tokyo Tech’s TSUBAME 2.0, connects each compute node to not one but two QDR Infiniband networks. Now that’s high performance!


About Marc Hamilton

Marc Hamilton – Vice President, Solutions Architecture and Engineering, NVIDIA. At NVIDIA, the Visual Computing Company, Marc leads the worldwide Solutions Architecture and Engineering team, responsible for working with NVIDIA’s customers and partners to deliver the world’s best end to end solutions for professional visualization and design, high performance computing, and big data analytics. Prior to NVIDIA, Marc worked in the Hyperscale Business Unit within HP’s Enterprise Group where he led the HPC team for the Americas region. Marc spent 16 years at Sun Microsystems in HPC and other sales and marketing executive management roles. Marc also worked at TRW developing HPC applications for the US aerospace and defense industry. He has published a number of technical articles and is the author of the book, “Software Development, Building Reliable Systems”. Marc holds a BS degree in Math and Computer Science from UCLA, an MS degree in Electrical Engineering from USC, and is a graduate of the UCLA Executive Management program.
This entry was posted in Uncategorized. Bookmark the permalink.