[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ih] Internet-history Digest, Vol 2, Issue 16

    > From: Clem Cole

    > But an open question in my mind crops up when you switch protocols at
    > the splice point ... By the previous definition these still meets the
    > second point of packets following through a networking splice point.
    > ... at one point have the two networks really become a single
    > interconnected network?

I think it really depends on at what layer the conversion from one to another
happens, and the resulting semantics. An application-layer gateway might
'unify' two networks for that application, but I don't think they they are "a
single interconnected network". E.g. we pretty quickly unified the email
networks, to present to the email users what appeared to be a single unified
network, but it wasn't really, it was just unified email.

Trying to do unification at lower layers can run into all sorts of
havoc. E.g. at one point the AI Lab guys tried to connect the PUP network at
MIT to the one at PARC; I no longer remember the details, but it didn't
utilize a single internetwork namespace, but translated addresses in packet
headers at the external boundary; so the inclusion of addresses _inside_
packets caused it to fail (horribly and spectacularly). Hence my focus on the
unified namespace for naming communicating entities.

IPv4 and IPv6 are closer, but for an IPv4 host to exchange packets with an
IPv6 host, one with an address which isn't expressible as an IPv4 address,
requires going through a NAT box - and the issues of addresses in the data

While leads me to perhaps another useful (necessary, but not sufficient)
measure of an 'internet' - the ability to deploy a new application between any
two arbitrary hosts with no changes needed except in them. Which does go back
to my original definition, though - the ability to get data from one to another,
through a single naming space. So perhaps this is more a corollary of having
an internet.

Internet-history mailing list
Internet-history at elists.isoc.org