Chelsio's 10-Gbit/s Ethernet Card Speeds TCP

Newcomer Ethernet specialist Chelsio Communications Inc. has introduced a card that can handle 10-Gbit Ethernet traffic at wire speed with 10-microsecond latency.

The PCI-X card, launched at the Grid Today 2004 conference in Philadelphia, is based on a single-engine ASIC dubbed the Terminator, with a TCP offload engine optimized for grid computing and data center applications.

The T110 card handles full TCP offload and can offer full-duplex performance at 10-Gbit/second speeds, the company said. It is designed for fiber short- and long-reach applications with XPak transceiver modules but also can handle coaxial interconnect utilizing Marvell Semiconductors' CX-4 interface. Chelsio's founders came from SGI's Protocol Engines Inc. subsidiary and were responsible for designing the SGI Origin supercomputer. Consequently, they understand the bottlenecks of using Ethernet framing as a channel-based alternative to interconnects like Infiniband. They also learned about the value of retaining standard Layer 4 protocols such as Transmission Control Protocol.

"You can develop a wonderful alternative technology, but if it doesn't fit the framework of standards, it doesn't gain a critical mass," said chief executive officer Kianoosh Naghshineh.

The team decided to offer a deliverable as a host bus adapter rather than a chip set, since most server and storage OEMs are familiar with HBA. Nevertheless, Chelsio says it has already entertained purchase requests for its Terminator ASIC. Terminator can handle 1 million concurrent TCP sessions, although the card is configured for no more than 64,000, indicating the future scalability of the architecture. Terminator has more than 5 million equivalent gates, implemented in a 0.13-micron process.Several large Ethernet players, including Intel Corp. and Broadcom Corp., have sampled 1-Gbit TCP offload engines, but Chelsio claims an additional benefit besides a 10x speed advantage. Chief technology officer Asgeir Eiriksson said that no aggregated RISC engines are used in the architecture. Instead, Terminator is based on a very long instruction word architecture in which no data is cached. Control plane CPUs hence show very low utilization rates when used with the T110 host bus adapter (less than 50 percent with a 2.2-GHz uniprocessing Opteron, less than 15 percent with a 1.5-GHz uniprocessor Itanium).

The board can achieve 7.8 Gbits/s when standard Ethernet frames are used, and it also supports the Jumbo Frames format. Eiriksson said the Terminator architecture allows better scaling at both the low and high ends. Multiple RISC engines can only obtain full wire-speed performance when 10 channels are used. The Chelsio board can achieve full speed with as few as two channels, but it does not run into cache thrashing when multiple channels are used, the company said.

At the Grid Today conference, Chelsio will demonstrate 7.8-Gbit/s performance with 9.5-microseconds application latency over Hewlett-Packard Co.'s Integrity and Opteron servers connected via a Fujitsu switch with fiber and copper interconnect.