|
|
|
|
|
Motleycast |
|
Grand Unified Theories |
|
Diversity in the Gene Pool? |
|
Algorithm and Data re-use? |
|
What we can do? |
|
What scary stuff is coming over the horizon? |
|
|
|
|
|
Motley crew of multicast… |
|
12 years of IDMR etc work has led to a lot of
multicast routing algorithms |
|
4 years of RMRG work has led to rather a lot of
reliable multicast protocols |
|
2 o 3 years of ffnet/yall work has led to a
number of application level multicast schemes |
|
Can we put it all together? |
|
|
|
|
|
Can we avoid folks re-inventing wheels? |
|
Tree construction |
|
Switch, router, server support for GRA |
|
Re-visit namespace |
|
Re-visit with security and billing |
|
Active Nodes? |
|
Key to Deployment: as many choices as possible! |
|
|
|
|
When you come up with a membership control,
access control, new tree building, maintenance, sub-tree, discovery, filter
algorithm: |
|
Think 3 levels – think link (shared v. NBMA);
think IP friendly; think application! |
|
|
|
|
Can we get synergy between different groups –
e.g. digital broadcast, and terrestrial ISPs; router vendors; application
and content service providers? |
|
New sub-effort within RMRG (Now that RMT and
PIM-SSM/bidir are everyday business spunoff)? |
|
|
|
|
Can we think incentives and see how to mix them
– as per qos, this is a multi-disciplinary activity – needs: |
|
Protocols & Information Theory… |
|
Economics |
|
Operations |
|
Stats/measurement/control/queueing etc |
|
|
|
|
New new thing is deployment of DWDM |
|
Optical cross connects go quite fast (e.g.
multiple OC-48 Nortel kit is commonplace now) |
|
Cannot do electronics to keep up and wavelength
assignment needs some control algorithms |
|
Ideal place to put multi-level motleynode. |
|
|
|
Notes
