Template Tools
Fifty Cents/Megabit for Big Bandwidth
Thursday, 30 June 2011 15:34
hunter_newby_matsumoto“You can buy a megabit today for fifty cents. In five years, it could be ten megabits for the same price.” Hunter Newby of Allied Fiber was answering questions at the Telecom Exchange event at Cipriani's. Fifty cents might be the price if you’re the size of an MSN buying 100 gigabits, maybe. For a small customer buying 1 gigabit, he said you could expect a price as low as a dollar or two at a major connection point, $5-10 at a regional hub. He doesn’t sell bandwidth but is in the middle of the market every day and well-informed. Asian and European prices are often 50% higher than the U.S.
      The continuing price drop is great for everyone except those selling bandwidth. It's why a service like Netflix is profitable when it would have failed five years ago. It allows carriers around the world to maintain unlimited bandwidth offerings and constantly raise the caps on others. It's why Julius never should have allowed AT&T to impose their 150 gigabyte cap a disguised price increase imposed to prevent full cordcutting.
      The drop in transit costs allows carriers to keep prices level despite the 25-40% growth in Internet traffic every year. 1 megabit/second is about right today for 10 retail broadband customers and is enough to minimize congestion. The transit cost is just part of the total bandwidth costs; the carrier also needs to upgrade switches, router and other gear inside their network. Cable companies with shared bandwidth will occasionally need to split nodes, a modest cost. Wireless is very different, because there's a limit on how much spectrum each tower can use.
      Here's how to convert from costs per continuous megabit to other contexts. These are based on industry rules of thumb, and vary based on traffic patterns, etc.
     $1/customer/month on a large cableco or telco. Large means you have fiber throughout your network and enough volume to use it efficiently. This cost hasn't changed in the better part of a decade. Equipment costs have come down about as fast as traffic has grown. Smaller carriers are often screwed.
      $2-5/customer/month for an efficient small telco lucky enough to cheaply connect to a regional hub. Anything below about 20,000 customers is seriously subscale with painfully high costs for bandwidth and many other needs.
      $5-20/customer/month if too small to buy bandwidth efficiently, especially if monopoly-like bandwidth pricing prevails locally.
       Large video providers, like Netflix or Amazon, are streaming movies at a cost of 2-4 cents. That allows a 2 megabit HD stream and some programming at 3.7 megabits. Small video providers pay much more. They buy from "content delivery networks" that cache on servers globally and provide other services.
    These prices do not apply to rural carriers who aren't lucky. Some pay $100 or $200 for that same megabit. Wireless costs are much higher, although much lower than most people think.
    High backhaul is the largest single factor holding back rural broadband. That became obvious to everyone at the broadband workshops and reported by the plan  Blair Levin of the plan and Sharon Gillett at the FCC have discussed doing something about this, but there's been no action yet.
Last Updated on Friday, 01 July 2011 04:25