[Iccrg] TANA comments

[Eddy, Wesley M. (GRC-RCN0)[VZ]]

The TANA BoF hasn't happened yet, and I'll be on a plane when it does, but I've looked at the problem statement and have a few short comments:
 
(1) I have no idea what an "advanced networking application" is.  I suggest actually defining this or finding a more precise term.  I assume you don't mean advanced in terms of age :).
 
(2) On "multiple connections" between 2 peers, this only has meaning when talking about TCP applications, with UDP it doesn't matter how many port pairs you use, the sending behavior can be the same.  The only difference is that with multiple UDP sockets you have multiple socket buffers to fill and the kernel may drain them in some different pattern than if you wrote all the packets to the same UDP socket.
 
(3) There is ambiguity about what transport services the targetted applications need; if it's datagram or reliable bitstream, either, both, etc.  It may not actually matter, but it might be useful to know in determining what work is actually needed here.
 
(4) The problem statement mentions that there is already a fielded solution.  Is the purpose of this activity to write up that solution, or to work on creating solutions?
 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://oakham.cs.ucl.ac.uk/pipermail/iccrg/attachments/20080728/048bda02/attachment.html