Every time I come across this concept I fall in love all over again, however as described, I don't think CCNx would be practical.
Firstly, it seems to require intermediary routers everywhere on the internet to keep track of all pending interests where they are between two endpoints. That's papered over a bit by aggregating like interests.
Secondly, while conceptually a beautifully unified scheme, the protocol described is actually pretty complex, with lots of optional bits, variable length addresses, strange routing schemes (describing routing policy in terms of a program running on an abstract VM!), baking security right into the core (which must be got right the first time around), etc.
Still. Perhaps in the future we might look forward to a simpler or more concrete system, similar in scope and design.
Every time I come across this concept I fall in love all over again, however as described, I don't think CCNx would be practical.
Firstly, it seems to require intermediary routers everywhere on the internet to keep track of all pending interests where they are between two endpoints. That's papered over a bit by aggregating like interests.
Secondly, while conceptually a beautifully unified scheme, the protocol described is actually pretty complex, with lots of optional bits, variable length addresses, strange routing schemes (describing routing policy in terms of a program running on an abstract VM!), baking security right into the core (which must be got right the first time around), etc.
Still. Perhaps in the future we might look forward to a simpler or more concrete system, similar in scope and design.