Re: initial issues

[Christian Schild <ipng@uni-muenster.de>]

On 15-Feb-2001 Ben Black wrote:
> I believe he means they only send traffic down a certain path (presumedly
> using some specific set of source addresses).  I'm not sure exactly why
> this would need special treatment.  Some clarification would be welcome.

Both of you are right.

Indeed the path selection can be done in the API by source address selection.
Nevertheless video streaming is a very obvious example, there _might_ 
be other cases, when the differentiation has to be done in the IP layer. 

As this is very vague and I can't think of a good example right now, I agree
to Itojun, that this should not be a shorthand task. 

Christian 
 
> On Thu, Feb 15, 2001 at 04:39:34PM +0900, Jun-ichiro itojun Hagino wrote:
>> 
>> >> goals of multihoming
>> >> - cope with L2 failures
>> >> - cope with upstream ISP failures
>> >> - load balancing, try to fill up two pipes we have toward upstream
>> >>       this needs more routing tricks.
>> >> - whatever you name it (but shouldn't dream too much, we need a
>> >> workable
>> >>   operational solution not a 20-year-to-deploy new protocol)
>> >What about QoS balancing?
>> >Consider a site with one cheap but narrow bandwidth upstream and a second
>> >expensive but wide bandwidth upstream. To save money such a site would
>> >want
>> >to use the wide one only if needed, e.g. videoconferencing and such, and
>> >the
>> >cheaper one for everything else. I'm not sure if this is covered by 'load
>> >balancing'.
>> 
>>      Do you mean looking up routing table based on TOS/traffic class
field?  
>>      I guess that was, in my mind, part of "more routing tricks" when I
>>      wrote the above text:-)
>> 
>> itojun