Re: [RRG] Taxonomy: 25 questions - Heiner Hummel's proposal

[Robin Whittle <rw@firstpr.com.au>]

Hi Heiner,

I have never been able to understand your proposal in any concrete
sense.  I understand you have some radical new way of using the
existing IPv4 address space.  If you are using some other addressing
system, then this means host changes and usage changes for the vast
majority of hosts, so the answer is Yes.

If by "ingress node" or "egress node" you mean "host":

> Robin, the Yes can be replaced by a No, if the ingress node added and
> the egress node removed the geographical coordinates. Of course, the
> ingress node must be enabled to add them. It must cache them, e.g. upon
> a preceding DNS lookup according to the experimental RFC 1712.
> I know RFC1712 was hardly been read: There where it nails down the
> encoding, longitudes and latitudes are terribly mixed up :-( 

then the answer to Q01 is Yes.

If by these you mean routers, then the answer to Q02 would be No,
because for your system to provide benefits, you are using a form of
addressing which can't be understood by routers or hosts in
non-upgraded networks.  So you can't maintain communications between
upgraded hosts (or non-upgraded hosts in upgraded networks) and
ordinary hosts - unless perhaps your hosts in upgraded networks can
also use an ordinary IPv4 address, in which case your proposal
couldn't provide benefits by getting hosts off conventionally
managed IPv4 addresses.

I am thinking about the transition time - not some time when
everyone runs your system, because that will never happen unless the
transition involves no loss of connectivity to non-upgraded hosts.

I suggest you write up a clear, diagrammatic, explanation of your
proposal.  If you can convince me it doesn't involve host upgrades,
and therefore only involves router upgrades, and if you can show it
provides significant benefits to early adoptors while also making a
significant improvement on the routing scaling problem, while
maintaining full compatibility with all non-upgraded hosts, I will
revise what I wrote.  Also, your proposal can't involve any option
headers etc. for packets transiting the DFZ, because the routers
there will not handle such packets at full speed - or perhaps at all.

  - Robin


--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg