A New Way to look at Networking
Why Hacker News recommends it
Recommended as the foundational talk on content-centric and named-data networking, and for its internet history stories.
What it is
Van Jacobson's 2006 Google Tech Talk argues for rethinking networking around named data rather than host-to-host connections. He gives the history and the unsolved problems of the current internet architecture.
Recommended by 3 HN commenters across 3 different months, 2012 to 2024.
What Hacker News said
Strong rec barrkel · 2024-09 · on “How the Hourglass Won”
Consider a society that has left Earth. If light takes minutes to travel between locations, is a chatty protocol what you want? Or do you need a whole host of different application-level solutions to this problem? There's a different lens to look at it. There's an old tech talk that got me out of the IP-shaped box a lot of us are trapped in: [this video] I got Gemini 1.5 Pro to summarize the transcript, which I butchered a bit to fit in a text box: The Problem: The Internet is treated as a binary concept: You're either connected or not, leading to issues with mobile devices and seamless connectivity. The focus on connections (TCP/IP conversations) hinders security: Security is tied to the pipe, n
Recommends seiji · 2012-07 · on “Wiki inventor Ward Cunningham develops federated wiki”
Ye olde, but still worth watching: Van Jacobson's Google Talk from 2006, "A New Way to look at Networking": [this video]
Recommends pololee · 2017-06 · on “Network Protocols”
Very good reading. I'd also recommend Van Jacobson's talk [this video] You'll learn some good stories about internet history. He has done a lot of work on TCP congestion control, especially the fast retransmission idea by using duplicate ACKs. There is an interesting thing about TCP. A lot of popular TCP implementation use city names, e.g. TCP Reno, TCP Vegas, TCP Westwood. TCP Westwood is a very interesting implementation. It has very intelligent way to estimate bandwidth, (not just based on duplicate ACKs). You may find this paper very interesting. [LINK]
Cites th-ai · 2016-10 · on “IPFS is the Distributed Web”
Any bashing on Named Data Networking? [LINK] Rather than the system getting less reliable exponentially as it scales up, it gets more reliable exponentially as it scales up. [this video] Integrity and trust are properties of the data; not of the way that you obtain it. [this video] There's no routing. You get rid of all of the routing chatter. You get rid of all of the routing brittleness. [LINK] We wanted to turn around the incentive structure where right now senders have a lot of power and receivers have basically nothing, they have to eat what they're fed. [LINK]
Cites pwdisswordfish2 · 2020-07 · on “Where Am I? NYTimes or Google?”
In what ways is this different/similar from "content centric networking"? [this video] (Google Tech Talk from Van Jacobsen on CCN many years ago)
Cites anaphor · 2021-10 · on “HTTP is obsolete – it's time for the distributed, permanent web (2015)”
Named-data networking is literally what this is describing, except NDN works as a replacement for TCP, not HTTP. In the long term I'd rather have it work at a lower level and not have to think about it much. [this video]
Cites xccx · 2021-10 · on “HTTP is obsolete – it's time for the distributed, permanent web (2015)”
NDN is the way p2p no huge IP infrastructure between nodes interest packets at narrow waist (like human attention, scarce) sign all packets, more secure, easier to trust|not hash for pointers, no surprises in containers crypto apply more cryptography, less web3 baloney broadcast radio native, no emulation of copper wire data, closer to where you want, find faster, keep content you want w/out intermediation, need to find "where" within IP/udp/tcp addresses [LINK] node to node [LINK] more application, less security hassle [LINK] hash name the data [this video] data integrity w/out need to trust foreign server