Ethernet exchange operator CENX is continuing to see rapid growth in participation by carriers and service providers. The latest milestone to be reached is 10 million Ethernet service locations, each of which is a unique physical address with a longitude and a latitude and all that. If it seems like just a short time ago that CENX announced 5 million accessible endpoints in its databases, well yes it was – two months to be exact. Obviously this number cannot continue to double every two quarters for very long without rapidly reaching all the locations worth reaching, but it is clear that uptake has been quite rapid.
The object of course is to bring Metcalfe’s law into play, i.e. that the value of a network is proportional to the square of the number of connected users. For Ethernet exchanges to really break through they need a certain level of critical mass, meaning that for carriers to use the system then when they want to reach an address there must be a reasonable likelihood that they will find a provider via the exchange that can reach it and will do so at a reasonable price. That’s why this 10M number is a useful one, as it clearly represents an appreciable percentage of the enterprise buildings any carrier might want to reach. Does anyone have an idea of a theoretical maximum for this number, whether domestic or international?
According to CENX’s CEO Nan Chen, the latest increase comes primarily from new service providers joining the exchange and adding the endpoints to which they can provide service, as opposed to expanded offerings from existing partners. CENX isn’t saying precisely who, but a quick look back over the past two months shows that perhaps a big part of the jump can likely be attributed to Covad and XO. Both bring a large number of potential endpoints to the table via the vast central office footprint they operate, reaching deep into metro areas via ILEC copper where necessary. They may not actually serve those endpoints yet, but they can and the whole point of joining the exchange is to make that fact more widely known across the industry.
While there is more room to grow in terms of connectivity, the next battle will be not where you can go but where people *do* go and how many – i.e. how many connections actually get made. However, it is still too early to expect that metric from CENX or any of the other Ethernet exchanges either operating or on the drawing board.
If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!
Categories: Ethernet
Join Our LinkedIn Group
Hi Rob. I think you’re conflating the number of discrete buildings connected to participating Ethernet exchanges at Layer 2 with the dynamic that is usually associated with the total number of end user nodes capable of any-to-any connectivity at Layer 3 on the Internet.
These are two different pairing classes, in terms of granularity and specificity, hence cannot be compared directly. Also, I hardly think that the 10 million number represents a measure of commercial “uptake”, but rather a market potential by virtue of the number of on-net buildings that exist arbitrarily within all exchange participants’ serving areas combined.
It bears remembering that Ethernet ‘connections’ at Layer 2, at this point in time, at least, are assumed to be customer NID to customer NID (trunk level, in other words), NOT end-user node to end-user node, as is the case today on the Internet.
Think of it this way: The atomic unit for Internet is the end user node; the atomic unit for an Ethernet exchange is an aggregation point within a building or billing entity, or the billing entity itself if there is only one NID involved.
These two models do not scale the same way, nor are their address ranges extensible beyond a certain point to the same degree, fwiw. If I’m mistaken about any of this, someone please speak up, cuz I’d much rather the limitations I’ve listed above not be the case.
frank@fttx.org