Fantastic video, better than many of the lectures I've sat through. If I could make a suggestion, though, I'd break a video like this up into three parts. Finding over an hour and half to watch a video is tough and RU-vid is set up to expect people to watch a the whole video in one go. Also, if you have any reading material you'd recommend I'd be interested.
Ahh, always good to see a new video, hopefully educational for people ! 3:55 yep, "smart endpoints, dumb network/pipes" 5:53 ohh, the irony of that 128 bits ! 22:27 not having 1 address was very useful in the times when protocols didn't have host/SNI headers. 23:52 I really hope CLAT just becomes common everywhere and be done with it. I hope Microsoft's recent IPv6 vulnerability didn't scare people. 30:53 strictly speaking, we should talk about organizations/AS-numbers, not just routers. 34:15 AND we send a notification ICMP-packet to the source. 44:12 bufferbloat PTSD... kicking in. 44:41 learned that the hard way. 45:20 I hope they allow ICMP ! 45:29 I know WebRTC uses it internally, kind of. But I've never heard QUIC ever tried to use it, if so, it most have been in the Google QUIC era, but pretty certain we knew by then it's dead on the public internet aka ossification, sometimes jokingly called OSIfication 1:00:36 which is why I'm looking forward to Multipath-QUIC, plug-in any IPv6 router with an active IPv6 Internet-link and increase your bandwidth or have automatic failover without loss of any existing connections 1:02:48 well, pretty certain that's not true, don't know about the specs, but certain hosts implement it with ARP (for example Redhat Enterprise Linux does it and Windows does it)
I skipped to the Multi-Path section because I'm familiar with the topics you discuss before it. I presumed it'd concern stuff like Multipath TCP, but in hindsight, that obviously wasn't gonna be the case since this is a video about IP, not layer 4. The Multicast stuff was interesting, as I wasn't familiar with PIM and although I had some vague ideas about stream delivery to many customers over multicast, your IPTV example really cleared things up for me. Thanks! To avoid confusing students, you might wanna check some of your example addresses/prefixes for routes. In particular, quite a few of your example /40s and /56s have non-zero bits beyond their prefix length, e.g. 2001:db8:420::/40 ought to be something like 2001:db8:4200::/40 or 2001:db8:400::/40 instead; and likewise for 2001:db8:420:56::/56, it ought to be something like 2001:db8:420:56::/64 or 2001:db8:420:5600::/56. I'm looking forward to your rundown of routing protocols like BGP, OSPF, and IS-IS.
I don't know how they managed to cockup IP v6, so badly. All they needed to do (not completely but mostly) was make v4 addresses wider and bamn, problems solved! Instead they fought like 4 year old, removing genuinely useful and adding useless features repeatedly. Now over 2 decades later, we're left with a confusing mishmash that works on some parts of the net and not others and conflicting "standards" isolating people like IP v4 never did! No wonder many consumers have no v6 connections, patchy v6 access at best and most none (or basically v6 over dodgy v4 at best).
