Categorizing Metro Fiber Operators by On-Net Depth

June 15th, 2011 by · 15 Comments

Following up on my update to the metro fiber provider and on-net building list, here is a color-coded table showing the ratio of metro route miles to number of buildings connected.  I’m not implying a race to the top (or bottom) of this table, just taking a cue from the numbers when it comes to understanding who does what and why across the sector:

The color choices are my own, I would welcome any suggestions for improvement.  Obviously some of the lines drawn are a bit arbitrary.

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: Metro fiber

Join the Discussion!

15 Comments So Far

  • Anonymous says:

    Would be great to see this for owned fiber only, and taking out the long haul as well. How much fiber does everyone own in the metro/access area – this is most relevant especially if you are comparing it to # buildings accessed.

    • en_ron_hubbard says:

      A few reactions:
      — why is “owned” compared to IRU’d significant?
      — this is described as “metro route miles” so I assume long haul is out of the calculation.
      — I would be very surprised if these numbers are accurate– Can Cogent only have one connection in every 9 route miles of metro? ABVT every three miles? If this were fiber miles it makes more sense surely.

      Am I wrong? Educate me.

      • ET says:

        Maybe because owned controls the sheath, IRU can’t always access at points they want without permission which isn’t always granted. Not totally sure on this.

        • Clevus says:

          In mostcases the sheath owner will not let you access the cable anywhere except at established splices or at the termination points

  • en_ron_hubbard says:


    In the information above is it clear to you that consistant definitions are being used by all reporting entities:
    — is it route miles or fiber miles? I assume fiber miles
    — is there any feeling for what counts as a “building” and is it only on-net buildings connected to fiber that are counted?
    What prompts this question is the seemingly HUGE disparity between two companies I am somewhat familiar with, TWTC and CCOI. Based on these numbers (builings/fiber mile) you would be forced to conclude that CCOI runs a pathetically inefficient network. But, as their financials amply demonstrate, they don’t. So something is badly askew. I would think the disconnect here is the definition of “a building”.

    This seems to make sense given that CCOI generates ~ $162,000 of revenue per defined building while the same TWTC number is $93,000 (2010 #’s). BTW, even though they have radically different numbers of claimed “buildings”, they each claim about the same number of discrete customers– ~26,000 each

    So, I think it is probably misleading to take one claimed metric, make an assumption that there is consistency in definitions across all the reporting companies, and reach any firm conclusions as to issues like “network efficiency” or “density”. Those concepts will show up in respective operating margins. So, for example, based on financial results alone my conclusion would be that XO (for example, which also has a large slug of Type ii connections) has a far less efficient and dense network/customer concentration than TWTC and CCOI, while the numbers in the chart above would suggest otherwise.

    All that said, the numbers are still interesting and this is the only place that kicks out such stuff. Thanks for compiling this information and putting it out. You operate a great site

    • Rob Powell says:

      Spoken like someone who rarely leaves Manhattan where a mile means something? 🙂 I assure you these are route miles, not fiber miles. AboveNet in its filings claims 2.1 million metro *fiber* miles!!

      As for Cogent and TWTC, you must understand that these companies have very different approaches. Cogent leases its metro routes from other providers, with a low fiber count per route, with the intent of hooking up only particular types of buildings – very big ones but with many small customers. TWTC has high fiber count, and hooks up much smaller buildings with fewer customers per building. They operate at a different layer in the ecosystem. The disparity you see relates to the difference in # of customers per building.

      As I said though, the table is not a race to the top or bottom, just a tool to help understand such differences – so your questions are welcome!

      You are right that ‘building’ is a poorly defined concept here and there can be consistency issues in the data. If I could, I’d collect more granular data by building type (LSO, multitenant office, wireless switching, datacenter, basic enterprise, wireless tower, etc). But it’s hard enough getting the level of data I already collect!

    • Rob Powell says:

      While XO’s business is less efficient in terms of where its revenue comes from, where they do have metro fiber it tends to be downtown in Tier I cities and more highly concentrated. TWTC on the other hand has many more miles in the suburbs and in Tier II cities that are less densely populated.

  • Rob Powell says:

    Perhaps to better characterize the sector I should also try to do a chart by the fiber mile as opposed to the route mile. This would show Cogent to be extremely ‘efficient’ in its usage of the fiber it has, while AboveNet would have massive amounts still unused. Shown in the context of the other tables, it might help clarify things further?

    • en_ron_hubbard says:

      Yes it would. And i do think it would facilitate a better appreciation of the respective busines models/efficiency/fiber rich and poor/etc., and answer questions that only wonks (like me) are interested in such as who generates most revenue per fiber mile?

      Btw, it was clear from the numbers that CCOI generates more revenue/building and has smaller average revenue/customer than TWTC, and while I eat dinner in Manhattan now and again I live a long way away from it. :0)

      One last thing re terminology– in some of its presentations TWTC calls them “route fiber miles” . I guess if a tower can be a building then that’s ok, but equally confusing. (Although if a tower can be a building can a building be a tower? I want to know.)

      • Rob Powell says:

        Actually yes, many downtown buildings are in fact towers, in that they rent space for wireless antenna on the roof. Basically, any man made structure that generates or consumes bandwidth is a ‘building’ for the purposes of this list – simply because it’s what I have been able to collect most consistently.

        • en_ron_hubbard says:

          Then that means fake trees are also buildings!! With that thought i will say goodnight.

          • en_ron_hubbard says:

            Actually, my closing thought is that a definition (‘Building”) that treats a tower that might have two or three tenants the same as a 1,000,000 ft2 multi tenant commercial office building is just plain wrong and misleading. If the metrics are wonky any stats derived are also wonky, but I guess that’s what we deal with.

  • Clevus says:

    This analysis has very little meaning in my mind. Enron’s comments on relative efficiency for a network in the metro environment is not a very good metric as the cost of a foo (or mile) of fiber is insignificant in comparison to the amount of data that can travel over it. A much more important thing to be looking at would be where thAt fiber is and how protected it is from future realignments. This coming from someone in the trenches so to speak on the fiber protection front. I am not sure how such relocation expenses are reported, but those costs as a function of route miles, or even revenus per mile might be a better metric

  • Brent says:

    I like the fact the TWTC is willing to build out if the ROI fits. We are getting them to build out to a single tenant building of ours with a 30mb commit, though it is part of a 4 site project though the other 3 are all off-net stuff. The pricing isnt crazy, but much better than we could hope for with anyone else, plus the fact its in qwest land theres not many good options.

Leave a Comment

You may Log In to post a comment, or fill in the form to post anonymously.

  • Ramblings’ Jobs

    Post a Job - Just $99/30days
  • Event Calendar