[arin-ppml] debunking the myth that Moore's law helps

[Michel Py]

> Scott Leibrand wrote:
> I've always looked at it from the supply side instead of
> the demand side (although both probably contribute).

That is a very valid point, and although I would have a few arguments
about the demand side driving the supply side, I will not start a
discussion about this because it would be as productive as the chicken
and egg problem. The relation between supply and demand is complex.

> The cost of transporting a bit is a function of the cost to route
> it (electronics), and the cost to transport it (optics). Routing
> costs seem to be tied to some version of Moore's law. Although
> there are specialized components like TCAM that don't observe the
> exact same factors as general purpose CPUs, they seem to be on a
> similar exponential curve.

Agree. Back to the original topic, can we agree that Moore's law does
not help building faster routers and might even have a negative side,
because the general purpose CPU may be on a faster version of Moore's
law than TCAM?


> Joel Jaeggli wrote:
> Actually WRT610n will route (as in make a layer-3 forwarding decision)
> about 300Mb/s it costs around $150, and yeah you can buy it a best
buy.

> Tyler Booth wrote:
> NAT is not routing.

I will point out that NATing is actually more work than routing (if the
routing table is small) as it has to replace addresses and port numbers,
and sometimes requires the ALG to de-encapsulate higher layers.

Joel, the WRT610n does NAT at 300Mb/s? I mean measured in the real
world, not only manufacturer's marketing?

Gee I'd like to have that kind of residential bandwidth for sure. I've
never seen anything that comes with a GigE interface yet. 

Michel.