From the Editor | Main Blog Index
September 10, 2009
It's probably no coincidence that at next Monday's High Performance Computing on Wall Street event, two of the conference headliners will be touting the joys of 10 Gigabit Ethernet for the financial digerati. Arista Networks founder Andy Bechtolsheim and BLADE Network Technologies CEO Vikram Mehta will explain why low latency 10 GigE technology has become so important to financial applications.
Since the demise of Woven Systems this past May, Arista and BLADE appear to be the only two remaining Ethernet vendors offering "ultra-low-latency" switching in a lossless fabric. The term "ultra" has no exact meaning, but since both vendors are touting sub-microsecond port-to-port latencies in their switch boxes, we can use that as an approximate definition. Sub-microsecond switch latency brings Ethernet into the InfiniBand realm, and makes this technology suitable for all sorts of HPC mischief.
Although the deployment of 10 GigE into HPC has been occurring in slow motion, 2009 may finally be the year it takes off. In July, Purdue University announced "Coates," a new 10 GigE supercomputer cluster that is expected to nab a top 50 spot on the next TOP500 list. And this week, Linux Magazine's Douglas Eadline outlined some reasons why he believes 10 GigE will start to gain significant traction in HPC over the 12 months.
My 10 GigE prediction is based on the following rule of thumb, Speed, Simplicity, Cost, pick any two. I believe 10 GigE will win because of simplicity and cost. IB is already faster and has better latency and if you need this level of performance you are not even looking at Ethernet direction. The joy of clustering is that one size does not fit all and you can build your cluster around your needs.
The expectation is that 10 GigE LAN-on-motherboard controllers will become commonplace on servers over the next 12 months and will tip the economic equilibrium toward the new technology. And thanks to faster switch ASICs and more mature software, lower-cost switches from companies like Arista and BLADE will also help to bring 10 GigE into mainstream service.
Today, both companies are offering top of rack switches in the $400 to $500 per port range. BLADE's Mehta recently told me they intend to be the first ones to break the $100 per port barrier, which would put it on par with high-end Ethernet switch pricing. And unlike Woven Systems, BLADE and Arista look to be at least reasonably financially stable.
BLADE, a private company, just announced its best quarterly revenues for Q309. According to Dell'Oro Group's Q209 Ethernet Switch Quarterly Report, the company is now the third largest supplier of 10 GigE switches, although Cisco still owns the lion's share of the market. (Purdue's Coates super mentioned above is based on Cisco and Chelsio network gear.) However, in the wake of Cisco's entrance into the server business, BLADE has aggressively gone after Cisco OEM partners, who now see the big network vendor as direct competition. Today BLADE's OEM relationships extend to IBM, HP, NEC and Verari Systems and are starting to pay dividends. The company says it has collected over 9,000 customers spread over 26 industry verticals.
BLADE also just completed $10 million in Series B funding, bringing on board three new investors: NEC, Juniper Networks, and an unnamed new investor, which Mehta characterized as a "big technology powerhouse." The investment from Juniper Networks is an interesting one, inasmuch as Juniper is also an Ethernet switch vendor. But the two companies have collaborated before -- although the exact nature of that collaboration remains somewhat of a mystery -- which could lead one to believe that these kids might get married someday, and present a more formidable threat to Cisco.
The younger Arista is not nearly as far along as BLADE, business-wise. On Dell'Oro's Ethernet switch report, the company doesn't even show up on the list, since presumably its market share is less than 0.1 percent, placing it in the "others" category. Nevertheless, the company does have $100 million in committed funding, which is presumably backed up by billionaire Bechtolsheim's giga-dollars. And since the company started shipping Ethernet switch gear in 2008, Arista has managed to collect around 100 customers, spread across HPC, manufacturing, health care and financial services.
But it's the latter vertical that has become one of the most interesting for 10 GigE vendors. Despite the fact that both Arista and BLADE have positioned themselves as cloud network providers (along with every other network vendor in existence), today the financial services space provides 30 percent of the revenue for each of these companies. A big chunk of this money is coming from the uber-profitable high frequency trading (HFT) segment, where the cloud computing model is entirely absent. Minimizing latency is at the heart of HFT, which is why both Arista and BLADE have been successful with this particular Wall Street application. In fact, some people just refer to HFT as low latency trading.
In our special coverage of the High Performance Computing on Wall Street conference next week, we'll delve more deeply into the importance of these companies in the HFT ecosystem and the financial services space in general.
Posted by Michael Feldman - September 10, 2009 @ 5:10 PM, Pacific Daylight Time
![]()
Michael Feldman is the editor of HPCwire.
No Recent Blog Comments
In a recent solicitation, the NSF laid out needs for furthering its scientific and engineering infrastructure with new tools to go beyond top performance, Having already delivered systems like Stampede and Blue Waters, they're turning an eye to solving data-intensive challenges. We spoke with the agency's Irene Qualters and Barry Schneider about..
Read more...
Large-scale, worldwide scientific initiatives rely on some cloud-based system to both coordinate efforts and manage computational efforts at peak times that cannot be contained within the combined in-house HPC resources. Last week at Google I/O, Brookhaven National Lab’s Sergey Panitkin discussed the role of the Google Compute Engine in providing computational support to ATLAS, a detector of high-energy particles at the Large Hadron Collider (LHC).
Read more...
The Xeon Phi coprocessor might be the new kid on the high performance block, but out of all first-rate kickers of the Intel tires, the Texas Advanced Computing Center (TACC) got the first real jab with its new top ten Stampede system.We talk with the center's Karl Schultz about the challenges of programming for Phi--but more specifically, the optimization...
Read more...
May 22, 2013 |
At some point in the not-too-distant future, building powerful, miniature computing systems will be considered a hobby for high schoolers, just as robotics or even Lego-building are today. That could be made possible through recent advancements made with the Raspberry Pi computers.
Read more...
May 16, 2013 |
When it comes to cloud, long distances mean unacceptably high latencies. Researchers from the University of Bonn in Germany examined those latency issues of doing CFD modeling in the cloud by utilizing a common CFD and its utilization in HPC instance types including both CPU and GPU cores of Amazon EC2.
Read more...
May 15, 2013 |
Supercomputers at the Department of Energy’s National Energy Research Scientific Computing Center (NERSC) have worked on important computational problems such as collapse of the atomic state, the optimization of chemical catalysts, and now modeling popping bubbles.
Read more...
May 10, 2013 |
Program provides cash awards up to $10,000 for the best open-source end-user applications deployed on 100G network.
Read more...
05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.
04/15/2013 | Bull | “50% of HPC users say their largest jobs scale to 120 cores or less.” How about yours? Are your codes ready to take advantage of today’s and tomorrow’s ultra-parallel HPC systems? Download this White Paper by Analysts Intersect360 Research to see what Bull and Intel’s Center for Excellence in Parallel Programming can do for your codes.
In this demonstration of SGI DMF ZeroWatt disk solution, Dr. Eng Lim Goh, SGI CTO, discusses a function of SGI DMF software to reduce costs and power consumption in an exascale (Big Data) storage datacenter.
The Cray CS300-AC cluster supercomputer offers energy efficient, air-cooled design based on modular, industry-standard platforms featuring the latest processor and network technologies and a wide range of datacenter cooling requirements.