The Future Internet as a Model for Future Public Systems

1
The Future Internet as a Model for Future Public
Systems

• Jon Crowcroft,
• Jon.crowcroft_at_cl.cam.ac.uk
• http//www.cl.cam.ac.uk/jac22

2
Future Internet Research

• Network research
• Resilience, scale, flexibility
• Content distribution
• Internet of things
• All optical core, all wireless access
• Etc etc etc
• Yada yada yada
• Blah blah blah

3
Future Internet Research

• Use as model for other publc services
• Past Internet delivered
• Decentralisation
• Federation
• Adaptation
• Evolution
• How?

4
Past Internet

• Decentralised, Hierarchical Routing
• Route around trouble
• Pool resources
• Distributed Resource Management
• Congestion control in end
• Pool resources
• Federation
• Policy route admits many business models
• Pool resources
• Layered Abstraction
• Can evolve independently
• Hence Web, VOIP, IPTV, OSNs, etc

5
Past Internet uses

• Advertisement/discovery
• Add a host, router net is trivial
• Rendezvous
• No pre-arrangement necessary
• Hierarchy
• Scale arbitrarily
• Names, addresses, routes, capacity etc

6
Architecture Design Rules

• End-to-end model
• Postel Principle
• Amdahl/Cray/Moores laws etc
• Consumer electronics scale

7
Three examples of applying model

• Transportation
• Use end2end model
• Energy
• Use federation model
• Government
• Use extreme decentralization model

8
Transportation

• The use of end2end principle-
• Separate routing and transport
• Person as packet, switched to destination
• Adapt routes in situ, en route
• Congestion control gt congestion exposure
• London was a good start -)

9
Energy

• Many many many providers
• Federate
• Build subtle meta-business rule system
• BGP
• Adapt (a la routing and traffic engineering)
• Use higher level information
• about demand
• Co-optimise Energy with other things
• E.g. 1. Transport 2. Interweb
• We have 6M EPSRC project looking at 2.

10
Government

• Decentralize subsidiarity
• IP name/address/content governance
• Hierarchy
• Hides complexity
• May be inflexible to multi-dimensional
• Social, economic, safety etc
• So my weakest example
• Typical geek trying to understand politics
• getting reductionist

11
Conclusions

• The lessons of the Internet are deep
• We can apply them in other domains
• Interesting synergies may emerge
• Might be a good FISB UK topic
• UK scale is excellent for ideas
• Much better than top down
• Database state unnecessary
• Avoid embarrassing errors like NPfIT

12
Questions