In a 2015 post discussing an OFC presentation of a well-known engineer in the business, titled “ROADMs Instead of Metro 100G Systems?”, the response from a technologist was, “Very interesting view, some may cry heresy.” fibeReality has been consistently rather bearish on public network deployment of metro 100G gear, including an earlier article: “Absurdity of Stressing Metro 100G Now,” posted in January 2015. During this month, we pointed to Cisco Systems’ silence on the matter as well as on our daily update page, noting Ciena stressing “the blur into both the convergence of long-haul and metro.” Although within the ill-advised endorsements of an optical supercycle, there is still a tendency to mention as one of the major contributing factors, high-end deployments of metro gear at service providers, the evidence continues to accumulate that the projections by some analysts concerning future equipment shipments have been extremely unrealistic.
Even when it comes to metro data center interconnections at 100G, there appears to be insufficient appreciation of the unique characteristics affecting the market size by some industry observers in that the distances can be much shorter than found with service providers, and there are fewer capabilities required than seen on a typical chassis. In addition, there can be substantial differences in the network buildout with the relatively few hyperscale operators that have a need for such a high data rate.
According to our Clash of Optical Component Vendors & Technologies in Data Center Networks report, Facebook’s infrastructure tends to be all long-haul in nature given its use of co-location facilities for picking up and dropping back local traffic. In contrast, our metro 100G study points out that Microsoft’s “interconnection distances can be as short as being across the street or on different floors of a building,” and therefore, the still considerably expensive coherent technology can be avoided.
When it comes to Verion’s metro sector, we believe that Ciena is being somewhat disingenuous when it recently stated, “in general I think the investment community got a little bit ahead of this” because the vendor appeared to be pushing this opportunity rather heavily itself to the Street. The fundamental problem is what we have referenced in an October 2015 piece, and it does not just apply to Verizon, but to AT&T and to CenturyLink as well. Deployment of metro 100G at these carriers continue to be limited “by a relatively small number of 100-gig client interfaces.”
Another important factor that is more general in nature was recently expressed by Viavi Solutions: “North America…, a year ago it was all vectors pointing in the same direction and all major service providers spending aggressively on access and metro infrastructure….This year, it seems like the switch was largely turned off, as we saw numerous M&A announcements come out from the major providers. It's literally within a week, a lot of the discretionary CapEx spend discussions came to an abrupt end.”
Of course, we have argued that much of this acquisition activity resulted from the push to bypass the FCC’s reach with the net neutrality rules. Despite the optimistic signs with a new leadership at the agency, Verizon, for example, may wait to see what really happens before it ultimately decides what to do with the Yahoo acquisition – which would ultimately have a significant impact on its equipment expenditures.
[written by Mark Lutkowitz]