Such overwhelming support for an MSA is rare. Some of Xenpak's appeal may be due to the fact that it's a very strict agreement. It specifies form, fit, and function for a networking link that's fully compliant with the proposed 10-gigabit Ethernet standard. In other words, the modules are required to be exactly the same size and do exactly the same thing, using the same printed circuit board connector and common electrical pin-outs.
In contrast, other MSA agreements out there are much looser arrangements, set up primarily to help manufacturers gain product acceptance by creating second sources for that product (see Infineon Gets 10-Gig Boost).
The inclusive nature of the Xenpack agreement is important because standardization is an important force in driving markets forward.
"It's a dramatic indication that 10-gigabit Ethernet is moving into the prime time," says Euan Livingston, marketing manager for Agilent's Fiber Optic Business Unit. "It is very much a mass market approach. It will make high-tech gigabit optics very user friendly, design friendly, and customer friendly."
The spec will simplify the process of building 10-gig systems and components. Systems players can design boards based on the Xenpak specifications, which will make them ready for modules when they become available. Chip makers can go ahead and design the chips that interface to the optical modules. And end-users will benefit from the flexibility -- they can buy a box and populate it with links on a pay-as-you grow basis, swap broken links easily, and upgrade links to different optics if they desire.
Other members of the Xenpak MSA are:
Xenpak sprung out of meetings of XGP, another 10-gig module standardization effort. When Agilent and Agere realized that a solution was nowhere in sight, they decided to jump-start the process by taking on the design themselves.
One of the problems was that the Institute of Electrical and Electronics Engineers Inc. (IEEE) had approved four different versions of the optics for 10-gigabit Ethernet modules (see IEEE Nails Down 10-Gig Ethernet Basics). Because different optics components vary in the amount of thermal heat dispersion, thermal performance was the main issue. Agilent and Agere had to design a link that worked regardless of the version of the optics inside the module. The other big design challenge at 10 Gbit/s was electromagnetic interference.
Whether the XGP group, which is developing a more compact connector, still comes up with its own module design remains to be seen. For now, the fact that so many vendors are united behind a single design should ease the adoption of the technology. Significant revenues from 10-gig Ethernet are not expected until 2003.
— Pauline Rigby, Senior Editor, Light Reading, http://www.lightreading.com