I've been doing a lot of listening lately about the state of IP networking and where things are headed. John Day on Programming Throwdown has been provocative, and Geoff Huston on the IPv6 Buzz Podcast also. Though progress is definitely being made, it's clear that IPv6 is the Python 3 of internetworking protocols: the switchover is taking DECADES. Some of the above blame this on IPv6's not providing much of use beyond a larger address space, an issue already mitigated using NAT.<p>It feels like things have been more-or-less the same since the original internet protocol came together in 1975. But when I learn about overlay networks such as TOR, IPFS, or Bitcoin's Lightning Network, I'm impressed by the innovation there. So what are the prospects for development of our global networking protocols?<p>I recently learned about a protocol called ILNP which splits the IP address into Identifier and Locator portions (IIUC). This way e.g. your phone could have the same Identifier no matter which location it occupies on the network topology, and somehow (I'm not clear on this) the mapping from Identifier to Locator is performed so packets can be routed. But development of that protocol, as well as its competitor LISP, seems to have stalled.<p>I guess I feel a bit disappointed and am looking for someone doing something exciting and radical in the networking world, not just on the overlay. It still seems to me like it will be an essential part of humanity's future, not just technically, but in terms of social structures, governments, etc. And yet things feel so stagnant.<p>What has you excited these days?<p>https://www.programmingthrowdown.com/2022/06/137-origins-of-internet-with-john-day.html
https://www.programmingthrowdown.com/2022/07/138-fixing-internet-with-john-day.html
https://packetpushers.net/podcast/ipv6-buzz-055-the-good-bad-and-ugly-of-ipv6-with-geoff-huston/
I'm excited about P2P/decentralized/distributed overlay networks. Still catching up so would be grateful for tips on resources.<p>Pinecone[0][1], newer initiative made by former Yggdrasil[2] maker(s).<p>CJDNS[3].<p>AIUI CJDNS relies on intermediary high-uptime discoverable router nodes which is what is motivating Pinecone. POKT[4][5] to CJDNS seems like what Filecoin is to IPFS.<p>I'm yet to get around to doing the groundwork of grokking more established solutions like B.A.T.M.A.N.,how all these pieces relate to each other and what's already possible. In particular what's realistically achievable on top of webRTC.<p>If you want to go real radical, well, Urbit.<p>[0]: <a href="https://fosdem.org/2022/schedule/event/matrix_p2p_pinecone/" rel="nofollow">https://fosdem.org/2022/schedule/event/matrix_p2p_pinecone/</a><p>[1]: <a href="https://github.com/matrix-org/pinecone" rel="nofollow">https://github.com/matrix-org/pinecone</a><p>[2]: <a href="https://yggdrasil-network.github.io/" rel="nofollow">https://yggdrasil-network.github.io/</a><p>[3]: <a href="https://github.com/cjdelisle/cjdns/" rel="nofollow">https://github.com/cjdelisle/cjdns/</a><p>[4]: <a href="https://www.pokt.network/" rel="nofollow">https://www.pokt.network/</a><p>[5]: <a href="https://piped.kavin.rocks/watch?v=-xgRUAA_p5E" rel="nofollow">https://piped.kavin.rocks/watch?v=-xgRUAA_p5E</a>
IMHO it’s pretty cool that you can stream HD video to the computer you carry in your pocket.<p>And if you need to connect the same computer to the NAS in your wardrobe to open a document or whatever, you can communicate securely using Wireguard/Tailscale.<p>Revolutions/upheavals don’t occur that frequently. But constant refinements can also yield awesome results.
What you want from networking is ubiquity, at the socket or wifi level, in the same way that what you want from a power socket is well-specified amount of power. So, you are not likely to get "exciting" changes in networking, but you may (and do) get backward-compatible upgrades in bit-rates.