The focus on the stated transmission bitrate is missing the point. It doesn't help whether the supposed transmission bitrate between two equipment ports is 10/100/1000...Gbps when an ever decreasing percentage of the bandwidth is used to carry revenue-driving client traffic.

In fact, in this era of dynamic, variable BW packet traffic, ANY static bitrate transmission channel based network is bound to suffer from very low utilization (in practice, less than 10% of available, statically provisioned, BW gets used for carrying revenue driving traffic) and/or from service/revenue affecting QoS problems due to congestions at network bottleneck points -- which will shift around over time, and occur, in statically provisioned networks, even while most of the network is technically overprovisioned (i.e. revenue-limiting congestions and cost wise wastefully low overall utilization can exist at the same time).

We need better utilization with congestion avoidance, not technically higher bandwidth rates. In reality, 40Gbps optics based network can support more revenue when applying appropriate dynamic capacity allocation techniques, than 100Gbps based network with conventional static BW transmission. Service provider's profit wise the difference in favor better utilization is even greater, given the lower capital and operating costs due to lower volume of capacity required to support the given revenues. 

I know that there is a significant amount of older fiber in networks that cannot handle 100G.

Mark wrote, "Our stance now is that we were at least somewhat incorrect on points 2-5."

Point 5 = "Tensile strength on existing fiber can apparently handle 100G."

Fiber and fiber optic cables have much higher tensile strength specifications.  You were correct and absolutely lowballing the tensile strength.  See TE's website.

http://www.te.com/catalog/Presentations/Fiber_Optic_cables.pdf

I like the idea of LR posting a Botched Predictions article.  That sounds like a commenting magnet.

Touché!  As with Light Reading, we do not bury analysis that we put out in past years.  At least we used the word, “probably.”  Our stance now is that we were at least somewhat incorrect on points 2-5.

re: "My main point is that the industry spends so much time on markets that have not materialized."

All tech-driven industries obsess on new technology.

Oh, and for a pragmatic view on why 100G will overtake 40G, check this out:

100G vs 40G

I had no idea Cisco created an artificial market for 40G.

:)

10G pricing has been reduced because of two big factors, pluggables and 2.5G.  Back in the early 2000's 2.5G was the optic of choice due to it's realtive price/bit and it wasn't as difficult to implement vs. 10G.  When the implementation concerns left (I.e. distortions, fiber issues) for 10G price was the only issue.  Most vendors back then wanted a 2-3x factor in pricing.  Then came SFF pluggables which changed the game against discretes in the mid 2000's; again this was a price reduction for both 2.5 and 10G.  40G may have missed it's adoption window (Qterra offered 40G in the early 2000's) and now 100G is getting the same SP pricing concerns.  

My best guess is 100G will dominate when the pricing concerns are alleivated and adoption will grow but it will mostly be a core adoption as 10G is, for now, more then adequate for Metro, Edge and NID.

Mark, a  year from now, one of us will be right ...

Sterling

Sterling,

We will have to agree to disagree on the pace of 100G.

My main point is that the industry spends so much time on markets that have not materialized.  We should be talking more about actual revenue coming from 40G.  

Actually, the more mature markets, like 10G, are always more interesting to analyze anyway.  What kind of minor refinements are we going to see?  When will the price drops flatten out?

Mark

Mark,

I don't necessarily read the installed base of 40G as a big negative for 100G. Systems suppliers have built 100G to work alongside 40G and even 10G channels. Many of the new DWDM systems were designed for 40/100G specifically. From operators, I hear alot about a cap and grow strategy for 40G and 100G - cap 40G, and then build new with 100G.

One other point: 40G never really got to a competitive price versus 10G. If that hasn't happened by now, I see that as bad news for 40G for 2012 and on. 10G takes the "low speed" links and 100G becomes the better high speed link option.

Sterling

Anyone notice that on Ericsson's Optical Networks Q4 Update that it had "reasonably strong 40G sales, especially with tier 1 customers, who have invested heavily in 40G infrastructure" -- and that it was up from the last quarter.

Mark Lutkowitz