QOS or more bandwidth

Sean M. Doran smd at clock.org
Tue May 29 14:23:36 UTC 2001

"Ukyo Kuonji" <kawaii_iinazuke at hotmail.com>

| The problem is, while most vendors support tagging and priority queuing, non 
| of the current vendors can support true end to end QoS. 

Where is the interoperability document (e.g. RFC) that 
describes "true" "end to end" QoS?

In the absence of such a document, from which anyone can build
an interoperable "true" "end to end" QoS system into his or her
product, I am tempted to believe that such buzzwords are weapons
in a DoS attack by marketroids.

|  you have no real way to ensure that the 
| high priority arrives at the destination in the same manner that it was 
| transmitted.  I'm talking about packet order, jitter, and latency. 

TDM.  What flavour would you like?  SONET/SDH?  PDH?  "Virtual dark fibre"?
ITU-Grid optics?

| Sure, it will probably get there, but will the data still be worth anything.

If you're not sure that it'll be worth anything on arrival, 
you shouldn't send in the first place.  Be liberal in what
you accept, and conservative in what you send is a good maxim.
RFC 2001 style congestion avoidance is an excellent example
of a system which follows this maxim closely.

| For Internet traffic, QoS/CoS is probably not worth it, as there is no way 
| to realistically do either across two or more network providers.

I keep waiting for a document which describes RSVP-TE/X.75 gatewaying
to solve this particular problem.

| Such applications will be Toll-quality 
| voice, production/broadcast-quality video, VPN, etc.

These are all travelling across multiple-provider paths today...

| The way I have seen it, either IP QoS will have to become a reality, or the 
| applications themselves will need to change to handle the poor CoS 
| substitute that is offered now.

The latter distributes the problem better.  The former requires
well-adapting, well-behaved, Internet-ready apps to subsidize the
migration of apps which are not all of these, and that will not fly.

(Note that even the poor CoS substitute is unnecessary for current
well-adapting, well-behaved, Internet-ready apps... that's one reason
they're so inexpensive.)


More information about the NANOG mailing list