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

[ih] vm vs. memory

On 10/24/2017 2:16 PM, Noel Chiappa wrote:
> I've thought extensively about how to do 'hard multi-homing' (as I'll name
> that particular flavour), and it's doable, but it needs to be built in up
> front.

In theory, it's not that difficult.  </insert ironic smiley>

Seriously:  insert not a smiley but a layer that looks like transport to 
the application but actually is session, and have the session open more 
than one transport connection that transit different paths.  (How to get 
the transports to do that has various 'solutions' as an implementation 
exercise or research topic, as you wish.)

With some tuning for synchronization and recovery -- and, ok, that's 
some engineering work, but not difficult -- the loss of one or another 
connection is merely reflected by a moment of delay to the application.

This requires participation by both ends, but the design and operation 
of this session layer wouldn't be that challenging.

BTW, TLS was really designed as a session layer, which means that 
modifying a few TLS libraries might make this a viable enhancement in a 
reasonable timeframe.

Dave Crocker
Brandenburg InternetWorking