ietf
[Top] [All Lists]

Re: draft-ietf-nat-protocol-complications-02.txt

2000-04-24 17:00:04
Sean Doran wrote:

Ohta-san:

| > No, not "or whatever" but "AND whatever".
|
| Do you mean "plug THEM in to your phone line and whatever"?

Yes, that is certainly one possibility.   I imagine the "inside"
interface would be some easy-to-wire LAN interface, enabling THEM
to communicate using whatever protocol/address seems appropriate inside.

I wonder how to decide the trade-off between the additional hardware
complexity of a multi-outside-port box and the additional software
involved in sharing state across several single-outside-port boxes.

Given the propensity of different hardware interfaces to spring up for
each new service, I'd argue having multiple boxes interacting, while
more complex, is more realistic.

Example: today I can get a DSL modem or router, and I can get a cable
modem. But the technologies of those interfaces seems to keep changing.
To build one device with multiple interfaces would require (due to the
changing technologies) a place to plug in interface modules. This raises
the cost of manufacture dramatically over building a box with a fixed
set of ports, so it's unrealistic to expect the market will favor it.

Sure, it's a harder problem. A good challenge for engineers to solve.

-- 
-----------------------------------------------------------------
Daniel Senie                                        dts(_at_)senie(_dot_)com
Amaranth Networks Inc.                    http://www.amaranth.com



<Prev in Thread] Current Thread [Next in Thread>