RE: [10GBASE-T] Economic feasibility or Market Potential?
I agree with much of what you say and I also agree with Pat's point
regarding the criteria and the "ROI" associated with the development
of a standard.
> Based on the above data, for the datacenter only:
> - 10GBase-CX4 will be applicable to ~25% of the links AT BEST.
> - 10GBase-T will be applicable to 95% of the links AT LEAST.
> Therefore, I have trouble reconciling in my mind the claim
> that #1 has broad market
> potential, while #2 does not.
I am not claiming that 10GBASE-T does not have broad market potential.
I am simply trying to understand what markets it solves, and how it compares
against available alternatives. Clearly, it does not compete with CX4 in the
>20m interconnect link space because of CX4's distance limitations, but then
it will have to compete in the <20m space against a mature (and most likely
integrated) PHY. It won't compete with 850nm for fiber links, but may lose
some of its market to 850nm for data center links.
I think the reconciliation might come if you include the other parts of the
market that don't rely on "Horizontal Links". CX4 is not intended to operate
over horizontal links and thus the data you are referring to is irrelevant
for evaluating its market potential.
I don't know if there is a study that shows how many "stacking links" are
being shipped per year (proprietary and standard) but that is a big part of
the market we are going after. In addition, we are looking for a standard
method of interconnecting switches and servers for short-haul applications.
Horizontal cabling is a different solution for a different space.
Hopefully, the questions raised lead to *answers* that help bring 10GBASE-T
to a better PAR and 5 criteria proposal.