[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Some questions on shim6
Victor,
Speaking for myself only:
Victor Grishchenko wrote:
Good day!
I am a newbie here; I am not an expert in shim and, frankly, I am not much
involved with IPv6. Still, I am interested in this topic mostly on
academic
grounds.
In particular, I have three questions.
1) Do you agree that shim6 is an "ad-hoc" (introduced to add multihoming
to IPv6)
I don't think it's ad hoc. There was a long analysis of many approaches
before shim6 was selected. It was carefully chosen to avoid damaging
existing functionality (in contrast to NAT, for example).
2) Do you agree to the metaphor "shim6 is a kind of NAT inside the
stack" ?
It's too simple. Shim6 is stateful, signalled, symmetric, reversible,
and adaptive address translation. Classical NAT is stateless, unsignalled,
asymmetric, irreversible, non-adaptive address translation that damages
existing functionality.
3) An alternate path of vectorizing TCP/UDP probably was considered in
the past. I mean, using multiple IPv6 addresses (a vector of) for a single
TCP connection; vectorizing all calculations performed by TCP. Do you have
any links to past proposals/discussions regarding this approach?
I think you need to study SCTP.
Brian