QoS - the most inter-disciplinary subject ever?
| Jon Crowcroft | |
| UCL, CS | |
| Jon@cs.ucl.ac.uk |
| What we tried: | ||
| QoS and IP – a subject with a great future behind it | ||
| Intserv,diffserv,measurement based admission control, price based admission control, mpls, traffic engineering, qos routing, red ecn etc etc etc. | ||
| Parekh’s thesis 8 years old | |
| RSVP 9 years old | |
| WFQ/SFQ/and Class of Service have been around for over a decade |
| Work in each area is ok, but is in isolation from the big picture – how can we fix this? |
Raise the entry bar to QoS proposals.
| There’s plenty of gold in “them there nets” | |
| Can we leverage some of it for QoS in the future? |
| Need to know: | ||
| Networks (telecom,ip,fr,atm,etc) | ||
| Control & Queueing theory | ||
| Economics&Law | ||
| Human Factors | ||
| Computer Systems | ||
| Physics… | ||
| Need to combine these expertise (c.f. Communications Systems Research Centers in Columbia, in Cambridge) | |
| Need to be open to adding more disciplines as needed | |
| Need to recognize non-problems (e.g. “over-provisioning” is often a reasonable solution). |
| Top slicing e-commerce…. |