Re: End-to-end (was Re: Non-Last Small IPv6 Fragments)

[Nick Hilliard <[e-mail]>]

Christian Huitema wrote on 15/01/2019 21:15:
> Actually, the pendulum went from "everything end-to-end" to "lots of
>  processing in the middle", and it is now swinging back. Developers
> have realized that if they want to keep a particular function
> "end-to-end", they should use encryption. For example, using HTTPS
> instead of HTTP prevents intermediaries from messing with the content
> of the pages. Using QUIC instead of TCP prevents intermediaries from
> messing with the transport headers.

it's more a change in what we mean by end-to-end. Historically it 
referred to transport layer end-to-end service:  device A could directly 
address device B, or vice versa.  Data integrity + confidentiality 
guaranteed by strong crypto seems like a more relevant proposition to 
today's internet though - it allows device A to talk to device B without 
having to worry about whether to trust all the devices in the middle, 
although it also generally dispensed with the symmetrical requirement 
for device B to be able to initiate communications with device A. 
Although orthogonal in approach, the functionality change is not 
necessarily a bad trade-off.

Nick