Re: [dtn-interest] Review for TCP CL

[Simon Perreault <[e-mail]>]

(adding dtn-interest to the CC: list)

Le 2012-08-28 11:41, Vint Cerf a écrit :
>     Third Paragraph
>     "...the TCPCL reside above the transport layer, i.e., at the
>     application layer."
>              Would you consider:
>     "...the TCPCL reside above the transport layer of the Internet
>     model, i.e., at the application layer."
>     I think the figure does a good job of showing the relationship; but
>     assigning BP to a 'layer' in the Internet model is always tricky and
>     a little extra verbosity might help.
>
> BP does not run at a fixed layer. It runs over pretty much anything that
> can carry framed bits. I think the best you can say here is that BP
> would run at application layer for the case that it (BP) is using TCP as
> its bearer with TCPCL as its convergence layer.

See my reply to Keith.

> Section 2.1 Definitions
>
> Fragment --  A fragment is a bundle whose payload contains a range of
>          bytes from another bundle's payload.
>
> I would have said "...whose payload contains a contiguous subset of
> bytes from another bundle's payload"

ok, fixed.

> Connection Initiator and Connection Acceptor
> There is such a thing as simul-init for TCP - could this happen with
> Bundle node that "know" there is a contact coming up and they both
> initiate the TCP connection?

You're absolutely right. This is important. Re-reading the text I 
realized that the distinction between initiator and acceptor are 
completely useless. There is no behaviour difference between initiator 
and acceptor. I will therefore remove those definitions and remove any 
reference to them in the text. Simul-init will consequently be 
implicitly supported.

Thanks!
Simon
-- 
DTN made easy, lean, and smart --> http://postellation.viagenie.ca
NAT64/DNS64 open-source        --> http://ecdysis.viagenie.ca
STUN/TURN server               --> http://numb.viagenie.ca