July 20, 2011
Intel has signed a definitive agreement to acquire switch silicon maker Fulcrum Microsystems. The deal adds Ethernet switch chips to Intel's growing portfolio of networking products. The agreement, whose terms remain under wraps, was announced on Tuesday.
Founded in 2000, Fulcrum is a privately-owned company that has been singularly focused on creating merchant silicon for high performance, lossless Ethernet networks. The company's two main product lines are its FocalPoint Ethernet chip family for network processing on switches and routers and its PivotPoint interconnect chip for packet processing on the line card. The foundational technology was developed at Caltech in the 1990s.
The emphasis on performance, and especially low latency, attracted performance-minded Ethernet switch vendors like Arista Networks and BLADE Network Technologies (now under IBM ownership). And until Mellanox scooped up Voltaire, the latter's Ethernet switches employed Fulcrum silicon as well. HPC storage maker Panasas also found a use for Fulcrum hardware in its ActiveStor product line. All told, Fulcrum claims its switch chips are used across 250 customer designs.
With cut-through latencies in the 200-300ns range and per-port power draw under one watt, Fulcrum's FocalPoint chip is just slightly less impressive than the latest InfiniBand hardware and on par with any competing Ethernet solution. For example, Mellanox claims its ambidextrous SwitchX silicon will deliver 160ns latency and about 2 watts per port for FDR (56 Gbps) InfiniBand, while providing 230ns latency and 0.6 watts when it switches to 10GbE.
The Fulcrum designs dovetail nicely with vendors who want to deliver cutting-edge 10GbE and 40GbE switching for high performance computing, cloud infrastructure, and other latency-sensitive datacenter applications. But the larger trend, and the one that Intel is likely to be most interested in long-term, is the commoditization of network switching.
Up until fairly recently, the accepted way to build switches was with custom silicon, often produced in-house by the switch manufacturers. With the appearance of merchant silicon vendors like Broadcom, Marvell and Fulcrum, that model is changing, threatening vendors like Cisco, which has long relied on its own chipsets. Just this week, it was reported that Lazard Capital analyst Daniel Amir sees Cisco beginning to transition to merchant silicon solutions.
That suggests that network switching is undergoing the same sort of transition that occurred in the server arena over the last 20 years, as most of the market moved from custom-build CPUs like Sparc and Power to x86. Over the last couple of years, Intel has seen its Xeon CPUs start to dominate in storage boxes and expects the same to happen in switches, replacing MIPS, Power CPUs and custom-built processors.
All of this reflects a larger strategy at Intel to establish a wider footprint in the datacenter. Since the chipmaker morphed its server division into the Data Center Group a couple years ago, the company has broadened its focus to encompass network and storage gear. The move parallels the shift of some of Intel's biggest enterprise customers, like IBM, HP, and Cisco, to package up "converged infrastructure" solutions, rather than just sell servers, storage and networking in a piecemeal fashion.
And although Intel has been selling Ethernet adapter chips and NICs on the host side for awhile, the Fulcrum addition will further their expansion on switch side, says Steve Schultz, director of marketing in Intel's LAN Access Division. "More and more we're seeing a lot of the switching and routing gear transitioning to Intel-based processors," he told HPCwire. "This is very complementary to that."
For Fulcrum customers, this is bound to be good news. Even though the company has been racking up impressive revenue numbers of late, expanding from $1.1 million in sales in 2006 to $6.4 million in 2009, it remains a small business with about a $100 million worth of VC money to answer to. Its five principal investors, New Enterprise Associates, Infinity Capital, Granite Ventures, Palomar Ventures and Worldview Technology Partners, will presumably all get their just desserts once the Intel deal closes.
In addition, being able to draw on Intel's considerable resources would put Fulcrum's product set on a much more stable foundation. And although Intel has not talked about moving the network chips into its own fabs -- the switch chips are currently manufactured by TSMC -- the chipmaker will undoubtedly want to leverage its world-class manufacturing technology.
Schultz says the existing Fulcrum team will remain intact and the business will be folded into the LAN Access Division, which itself is run out of the Data Center Group. The acquisition is subject to shareholder and regulatory approval, but according to Schultz no glitches are foreseen. The deal is expected to close in 30 to 35 days.
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...
May 09, 2013 |
The Japanese government has revealed its plans to best its previous K Computer efforts with what they hope will be the first exascale system...
Read more...
May 08, 2013 |
For engineers looking to leverage high-performance computing, the accessibility of a cloud-based approach is a powerful draw, but there are costs that may not be readily apparent.
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.