[arin-ppml] Ending point to point links as a justification for a /30?

[Owen DeLong]

On Jul 30, 2010, at 6:43 AM, <michael.dillon at bt.com> wrote:

>> Express a 128 bit number in 32 bits such that all possible 128 bit
>> values
>> are uniquely expressed in the 32 bit field.
> 
> This problem was solved a long, long time ago at least as early as
> 1930 when BAUDOT code was enhanced by the ITU's predecessor to become
> International Telegraphy Alphabet No. 2. 
> 
So every BAUDOT system with zero software modification was able to
send and receive ITA No. 2 symbology? That's not what I remember
reading.

> A more modern example is the Unicode encoding known as UTF-8 where
> escape codes followed by multiple bytes are used to represent codes
> which would not otherwise fit into an 8-bit byte.
> 
I don't know of a single ASCII-only system that did not require software
updates to be able to accept user input or display UTF-8 characters
outside the normal ASCII range.

> An IPv4 extension might take over the class E address space (E for
> Escape and Encapsulation) to use as Escape and Encap addresses for
> addressing portions of the extended 64-bit IPv4 address range that
> are not accessible with straight 32-bit addresses. Or you could

Even if you did that, you would need software updates on EVERY
end system and to every piece of client/server software in order
to support it. It would not be backwards compatible.

Bzzt... But thank you for playing.
> 
> 
>> If you solve this problem, please contact me, I'm sure we can make
>> money
>> from your solution.
> 
> People make money by selling the Brooklyn Bridge, but that doesn't mean
> it is a good thing to get involved in.
> 
If you can solve the problem in a way that does not require software
modifications on the existing IPv4 hosts, I'm pretty sure that would
do well selling in today's environment. However, you can't actually
do so. Each of the techniques you described above did not pack
the unique 128 bit values into 32 bits, it merely provided a way to
flag the 32 bits for "but wait, there's more". That's not a solution,
that's a side-step.

Believe me, if you can find a way to represent 340 undecillion
unique values in 32 bits and make it actually work 100%,
I guarantee you there's legitimate money to be made.

Owen

P.S. Rather than working on an IPv6 replacement, I think we should
look at resolving the routing paradigm so that we can develop a
scalable IDR system. I believe that will require modification to the
packet headers, so, it will be an IPv6 replacement on one level, but,
this particular system should be able to be deployed in a backwards
compatible fashion much like the 32-bit ASN setup.