RE: (SSIG) Taking the winning route
It is very easy for an advocate of any particular technology to present a
crude "cost analysis" which indicates that his or her approach is the least
expensive, and all of the others are more expensive. At the March 1999 call
for interest, one such "cost analysis" was presented by an advocate of the
850-nm serial/new MMF approach. At the time, I pointed out that the
presenter's assumptions and conclusions were completely different than my
own, and that, in fact, my own analysis led to the opposite conclusion.
The truth is, that until a real product is manufactured, and all of the
labor, yield, and volume issues are well understood, it is not very useful
to argue about cost. I could argue that a quad IC and 4 lasers running at
1/4 the speed, using packaging technology suitable for 3.125-Gb/s are less
expensive than a single laser and a single IC running 4 times faster in a
high-speed package. Someone else could argue that the high speed technology
is no more expensive than the low-speed technology and that the 4-channel
approach is 4-times as expensive! Until we make and sell our products to
customers, the debate can not be settled.
The IEEE 802.3ae is not an appropriate forum for quoting prices or sharing
cost models, delivery times, etc. This is confidential information to be
shared between a vendor and a customer. It is the customers, after
collecting all of the relevant information from potential vendors, who will
ultimately decide what is implemented and what is not. This reflector is
full of PMD vendors trying to argue, without data, why their particular
approach is the best. Fortunately for all of us, there are a few PMD
customers who are involved in the standards process as well. The success of
their businesses depends upon them being able to sort through all of the
data (both publicly and privately presented), and determining the solution
or solutions that best suit their needs.
-Brian Lemoff
Agilent Labs