[HSSG] Economic and Technical Feasibility
Following John's encouragement that we use the reflector for consensus building, I would like to share my current concerns, based on what was shown so far, regarding economic and technical feasibility.
1. In the CFI presentation, Photonic Integrated Circuit technology was featured as proof point for the economic and technical feasibility of a multiwavelength PMD. Recent work in the HSSG / Fiber Optic Ad Hoc however is based on traditional implementation technologies. While I believe that technically it is feasible to build transceivers with the technologies discussed, I am concerned that these will be too expensive. It would be helpful to hear from a lot more vendors who plan to invest in 100G transceivers and from vendors who plan to make components (e.g. 25G PIN diodes, 2x10G -> 20G SERDES, state of SiGe market, etc) needed for building these transceivers. Those of you who lived through the introduction of 40G into the market know how painful that was in terms of finding enough suppliers and costs.
2. 100G MAC feasibility
We had excellent presentations at the Dallas meeting regarding 100G MAC feasibility and a proposal for a 10x10G interface (CTBI?). After thinking about the material, it seems to me that some important topics have not been addressed:
a. Is there any industry experience with 10 X serial 10G in one package? It would seem to me that this would come with serious signal integrity challenges both on the chip and its package as well as on the PCB. Can the experts bring some material to the group and convince us that this is feasible?
b. What is the state of the art regarding 10G serial I/O on CMOS chips? Is this bening done at 90nm or does it require smaller geometries? If not currently available, when will it be available?
My apologies if the answers to all of my questions and concerns are obvious and well known to all.
----- Original Message -----
From: "John DAmbrosia"
Subject: [HSSG] BER Issue
Date: Tue, 19 Dec 2006 20:26:53 -0800
The HSSG reflector has not been highly utilized over the past couple of months. I would like to encourage all to take advantage of this medium for consensus building.
As I noted in my other email to the reflector, we have not chosen a BER objective for this project. There was considerable discussion on the subject back at the end of August, but no decision was reached.
In my review of the responses, I have seen some pushing for a BER of 10^-15, while others feel that this is impractical and will be expensive. In private conversations, I have heard some suggesting a middle ground BER of 10^-13.
I encourage individuals to consider this issue and comment further here, as well as consider presentations regarding this matter at the January Interim. It clearly has the possibility of impacting our responses to broad market potential, technical feasibility, and economic feasibility.
Menachem E. Abraham
Columbus Advisors, LLC
20 Park Plaza, Suite 909
Boston, MA 02116