NANOG,
I enjoyed Yakov's presentation on charging for network route
announcements. However, I seem to have a rather significant
problem with it.
According to Yakov, increasing the cost of network routing
announcements will decrease the overall number of routes, and
increase aggregation. Introducing dollar-fixed cost to network
announcements would provide a fair and equitable system in which
desire would not run rampant, but be tied to a fixed sum, ie
money.
In a standard capitalistic model, when supply is fixed, cost will
create an inverse relationship with demand. However, the 'supply'
of network routes is not fixed.
One could make a good point that it is fixed by saying 'Oh yes,
for so many dollars routers can only handle X number of routes in
their table optimally'. If we were to buy this (which I actually
do) then we would ask 'Who is it that determines what percentage
each provider/nsp gets of those X routes?'
If there were an autonomous 'Internet Company' that 'owned' the
abilities of the Internet, and could claim ownership of 'routing'
then they could determine that routing was worth 1 Frobozz. This
would determine that a 'valid' router could handle X number of routes,
therefore each route cost 1/X Frobozz.
However, we've not such a standard, nor a body to enforce it.
Akin to Lothberg's argument, we've not a supply, governed by an
entity, or set of entities.
While the routing tables of the internet connections points are
analagous to gold in many ways, in supply and control, they are
not.
How would this be enforced in today's system?
-alan
I enjoyed Yakov's presentation on charging for network route
announcements. However, I seem to have a rather significant
problem with it.
According to Yakov, increasing the cost of network routing
announcements will decrease the overall number of routes, and
increase aggregation. Introducing dollar-fixed cost to network
announcements would provide a fair and equitable system in which
desire would not run rampant, but be tied to a fixed sum, ie
money.
In a standard capitalistic model, when supply is fixed, cost will
create an inverse relationship with demand. However, the 'supply'
of network routes is not fixed.
One could make a good point that it is fixed by saying 'Oh yes,
for so many dollars routers can only handle X number of routes in
their table optimally'. If we were to buy this (which I actually
do) then we would ask 'Who is it that determines what percentage
each provider/nsp gets of those X routes?'
If there were an autonomous 'Internet Company' that 'owned' the
abilities of the Internet, and could claim ownership of 'routing'
then they could determine that routing was worth 1 Frobozz. This
would determine that a 'valid' router could handle X number of routes,
therefore each route cost 1/X Frobozz.
However, we've not such a standard, nor a body to enforce it.
Akin to Lothberg's argument, we've not a supply, governed by an
entity, or set of entities.
While the routing tables of the internet connections points are
analagous to gold in many ways, in supply and control, they are
not.
How would this be enforced in today's system?
-alan