Mike, If I follow your observation, routing computation growth is non-linearly related to number of routes. Or is it orthogonal? If the growth is related to the announcement, this infers that costs incurred for the announcement of routes by ISP FOO to ISP BAR would be non-linear with the quantity, correct? If those costs are non-zero, then BAR needs to be compensated by FOO for the announcement. One way for BAR to recoup its costs would be for it to collect a payment from FOO. Thus FOO "buys" a routing slot from BAR. From your observation, these payments may be non-linear related to quantity. Note that BAR may owe a similar payment to FOO for consumption of the same resources on FOO's network. Without the payments you have market failure -- consumption of the resource is unchecked ("tragedy of the commons") and people who "want" or "need" to make a route announcement, and who *could pay* for the resources that they consume, cannot do so. Thus the REAL problem is that we don't have markets for the determination and allocation of these scarce resources whether you call them "routing slots" or "routing complexity". Redirected to piara. -scott At 12:09 PM 3/2/97 -0500, Mike O'Dell wrote:

Yet again people keep talking about "the size of the routing tables" as being the deep problem, and this makes people say silly things like "FOO is protecting router memory".

Thinking about it this way is funamentally and fatally incorrect.

The REAL problem is the growing complexity of the ROUTING COMPUTATION, not the size of the resulting forwarding table. even if routers had infinite memory, we would still be crushed by the routing computation if allowed to grow unchecked.

-mo