> I'm not sure that the reasoning behind this actually holds. The
> transit (TIER1) provider have invested in a network. That network
> already have a fixed cost. That means that the provider need to charge
You are assuming that you activate everything from day one.
> for that ,+ 100% (or another huge percentage), to be able to
> survive. The write off (don't know the name of that in English,
> Swedish: avskrivning) is at least 15 years for international
Depreciation. Actually, how you calculate this for a optical network is
a science in itself. Fibers, ducts, civil-engineering etc is
depreciated on 15-25 years, depending on type of links. Long-haul sea
cables are normally shorter. Active equipment is depreciated on 3-5
years, normally five years.
As you activate more channels, you will move more cost to depreciation
so your cost per bit will be fairly stable in the beginning of the life
of the system, and later decrease fast. This is more or less linear to
the utilization of the system.
> networks. So the customer is not paying for the amount of traffic in
> the network but rather the investment. So paying for traffic is just
> wrong (you could read bullshit if you like but I wouldn't dare to
> write such foul language :-). It's all about maximizing income and
> minimizing cost and getting the customer to belive "I'm just paying
> for what I use".
This is not true. This is the QoS trap. "Let's sell the customers
bandwidth at price X and then see if we have enough money to actually
build a network that can handle it". It's apparently abbreviated
- kurtis -