Re: [rrg] Host changes vs. network changes

I try to be very careful when telling other folks (in this case the 
transport and application folks) what they want.

While the multi-path TCP is definitely of particular interest to hosts 
with direct connections to disparate networks, I do not think that is 
the only place it is of interest.  Various companies have made 
significant businesses out of the differences in connectivity in the 
wired networks.  And letting multi-path TCP use multiple of such paths 
seems to actually make the clients more cooperative with the network, 
while getting them better performance.

Hence, I am concerned that we are hiding the current target of interest 
from the folks whoa are interested.  Should they care?  I am certainly 
not in a position to say no.  It is possible that this is essentially 
equivalent to the necessary tradeoffs for aggregation, where the routers 
in the inter-domain sense do not have visibility to all of the details 
(caveat Nimrod, and being able to guess when one needs the details, but 
that is a different debate.)  But it does not appear to be the same 
level of mandatory trade-off that is implicit in aggregation.

yours,
Joel

Scott Brim wrote:
...
> Can future routing and addressing scaling can be easily engineered in a
> way that also allows endpoints easy control over what they want?
> 
> I think the situation is not as bad as it seems from your portrayal.
> What endpoints want is essentially: control over cost, robustness of
> connectivity, and good throughput, and IMHO the most urgent scenario for
> multipath is when there are significant differences in access network
> behavior -- when the endpoint itself is connected to multiple networks.
>  The situation you're describing, where the access network itself is
> multiply connected, is important to be able to multipath in, but not
> urgent, and ideas are being tossed around about choosing exit points
> through collaboration between the endpoint and the network, etc.  In an
> encapsulated world an endpoint would be able to choose the exit path
> from its provider but not the entry point into its correspondent's
> provider.  That may be tolerable.  So if there is a problem, it may not
> be serious.
> 
> Scott
> _______________________________________________
> rrg mailing list
> rrg@irtf.org
> http://www.irtf.org/mailman/listinfo/rrg
> 