A new way to look at networking by Van Jacobson

14th February 2021 at 10:41am

A new way to look at networking by Van Jacobson

Originally found on Bret Victor's What can a technologist do about climate change? (November 2015)

This presentation provides a nice overview of the context when TCP/IP was invented and what is the solution to replace it.

One aspect Van Jacobson often insists on is that you need to have the right point of view, you need to know what is the problem if you want to be able to solve it. In my opinion, it's a sign of a great problem solver.

TCP/IP was invented at a time where there wasn't a lot of data nor a lot of machines. It was mostly of lot of users per machine.

But now the world is different. There are a lot of machines per user (phone, personal computer, work computer, etc.). TCP/IP was extremely good for the problem it solved at the time but it was was not invented for the current world.

TCP/IP was built for a conversation model between two entities (e.g Phone A <-> Phone B). But the internet doesn't do communication. More than 99% of the traffic is point to multi-points (e.g multiple users ask for one URL and get the data). What we need now is a dissemination model, not a communication model anymore.

One of the thing that annoys Van Jacobson the most is that Internet is a binary thing: you are connected or you are not, you have information or you don't, there is no middle ground.

The design philosophy of the new technology to replace TCP/IP:

  • Data got a name, not a location
  • Integrity and trust are derived from the data, not the channel (think crypto hash)
  • Anything that moves bits in time or space can and will be used to communicate (anything! disk, memory, wires, etc. – I guess it means that my disk can serve as a "network" between my apps on my computer)

His new protocol can be found on Named Data Networking.

Related to