There is a lot of interesting history and things to get you thinking in his presentation.
Pay attention to the end of the presentation for this:
Wireless was absolutely a part of the internet architecture.
I'm uncomfortable with the idea that the emergency network is dependent on a system that is often know not to work during emergencies.
Or If the emergency occurs in an area where there isn't any cellular infrastructure. (Base-stations, back haul)
So I actually like the idea of doing down a different path.
Which is a mesh like thing, that can be ad-hoc.
I also rather like the idea that everything transported in that system is IP based. The part that I like is, it doesn't care how its carried, it doesn't know what its carrying. It could be voice, it could be video, the network is oblivious.
By the way the police and fire departments and EMS today have a lot of trouble communicating. Because their radios have to work at RF. If Instead if we said the whole base is IP.
And if you need to communicate by voice we will use VOIP. And if we have to we will do RF transformations in some portable nodes.
The whole idea is that your compatibility on an end to end to based on VOIP, and not on radio compatibility. We'd have a much more general architecture.
For some reason that unknown reason that obvious fact has been resisted or ignored by the public safety environment.
Thanks for including this interesting talk by Dr. Cerf. I appreciate you putting it on your blog.
Well, it's interesting, but in truth TCP/IP is entirely incompatible with dynamic, intermittent connections. The routing algorithms don't comprehend mesh networks, DNS algorithms don't handle rapidly changing IP addresses, and interruptions trigger prompt disconnects. We can patch around these problems, but it's a shame Cerf and his colleagues didn't fix them architecturally from the beginning.
Post a Comment