Friday, July 20, 2007

40 and 100G in Ethernet's future

Kudos to the engineers who toughed out a hard year at the IEEE's Higher Speed Study Group to hammer out a consensus accepted this week by the broader body that the IEEE will write standards for both 40 and 100 Gbit/s versions of Ethernet. Their work will pave a long and rich road map for years to come that should serve both data centers as well as central offices.

Debates over whether a 40 Gbit/s spec would disrupt the networking world, and prospects for compromise looked pretty dark even a week ago.

Hard as the work has been to date, the big job starts when a task force is formally convened. The effort to define a 10 Gbit/s version of Ethernet to run over copper cables was tough going and commercial products based on it still face plenty of design and cost hurdles. This week's decisions promise 100 Gbits/s over 10 meters of copper cable, and that will require some significant engineering. So congrats, and roll up your sleeves.

3 comments:

Anonymous said...

The inclusion of 40GbE is likely to undercut 100GbE in the marketplace. 40GbE will be built on 10G optics. There will be lots of packaging effects to work through for 40GbE but 4x10G optics should be much cheaper than 4x25G optics. It's hard to think of why anyone in the NPU/TM through optics supply chain would engage in the development of 4x25G, particularly if 40GbE looks like it will take the legs out from under the 100GbE market. Good luck to the 100GbE-keen switch vendors like Cisco and Infinera in trying to source components for 100GbE.

One very interesting consequence of including 40GbE is the likely appearance of a 40GbE WAN standard based on OTN transport.

Anonymous said...

While the discussions within the IEEE so far have been focused on clearly identifying what the market needs the next step will be to hammer out a standard.

Having customers, the real consumers of the technology come in and articulate their needs and plans for the future is greatly appreciated and absolutely necessary. But that's the easiest part. Now the technical experts take over and the engineering work begins. That requires significant commitment from companies and investors to provide the finanical support to not only send the companys top experts to these meetings for the next 18 to 36 months but to commit funding for product development. These very, very expensive programs.

Now consider 10Gbps for a moment. How many module vendors, subassembly vendors, and silicon vendors, etc have realized a return on their 10Gbps investments? As bleak a picture as that paints, now consider that those same business managers, VCs and others will be asked to fund the R&D for either 40G or 100G, or horror of horrors, both 40G & 100G.

Capital investment always seeks the best return possible. How will the market opportunities for 40G & 100G products stack up against the ROI potential of other investment opportunities?

Anonymous said...

The challenge to the success of any technology is maintaining focus while keeping an eye on the overall cost equations (there many different equations based on which dimension of the problem space is being addressed. This is why 10 GbE has largely failed to garner the vaunted volumes and low cost that so many have predicted.

The success or perhaps tepid success of 40 and 100 will be determined not by the technologists and their ability to create specifications but by technologists maintaining focus and developing specifications that do not penalize the volume usage in order to satisfy the other usage models who which may have more lattitude in their cost models. The failure to ramp to volume in the 10 GbE space is largely due to too many IEEE members putting forth numerous requirements that are cost adders to the volume space - today, the only real volume is products targeted on short-distance data center topologies. If 40 and 100 are to have any chance to succeed, then the IEEE must not impose these cost adders on the volume offering which again will be in the data center not the metro or wan configurations. The technologists need to stay focused on creating viable specifications that place the cost only on those products that will actually use a given feature or capability while insuring that the volume applications are as lean as possible. No one wants a repeat of the 10 GbE fiasco which led to more specifications being created to bless what people really wanted to deploy rather than what some companies wanted people to deploy so their costs were spread across all rather than those who used that particular flavor.

Let's hope the technologists will get it right this time and not create a tepid lose-lose situation as they have done in the past.

 
interconnects