Is that Ronald behind the purple mask?
", "time": "2022-07-29T14:03:27Z"}, {"author": "Juliusz Chroboczek", "text": "(Since he pronounced my name correctly, it's probably him.)
", "time": "2022-07-29T14:04:07Z"}, {"author": "Juliusz Chroboczek", "text": "Does anyone have a link to the MANET mailing list posting Ronald referred to?
", "time": "2022-07-29T14:21:45Z"}, {"author": "Juliusz Chroboczek", "text": "Everybody hates raw sockets!
", "time": "2022-07-29T14:23:16Z"}, {"author": "Juliusz Chroboczek", "text": "How many pages is LTE?
", "time": "2022-07-29T14:44:05Z"}, {"author": "Donald Eastlake", "text": "Error in my presentation: slide 15 \"<SA, Sequence Number>\" should be \"<MeshSA, Sequence Number>\".
", "time": "2022-07-29T15:03:57Z"}, {"author": "Juliusz Chroboczek", "text": "Agree with everything David said, except perhaps the point about assigning destinations at the source being a bad idea.
", "time": "2022-07-29T15:15:52Z"}, {"author": "Zheng Zhang", "text": "We can just set the all bits in BitString to 1
", "time": "2022-07-29T15:21:50Z"}, {"author": "Juliusz Chroboczek", "text": "But we still need to assign an index to each router, so it knows which bit to reset when it forwards a packet and it knows who to forward the packet to.
", "time": "2022-07-29T15:22:33Z"}, {"author": "Juliusz Chroboczek", "text": "Meetecho, I'm seeing a couple of seconds of latency in the chat.
", "time": "2022-07-29T15:23:05Z"}, {"author": "Zheng Zhang", "text": "only edge routers needs to be assigned. And we can use optimized method to assign the same bit for all the receivers if it's needed
", "time": "2022-07-29T15:23:21Z"}, {"author": "Lorenzo Miniero", "text": "@Juliusz Chroboczek do you mean latency as in time it takes to see a message you write being echoed back?
", "time": "2022-07-29T15:24:16Z"}, {"author": "Juliusz Chroboczek", "text": "That's interesting, although I'm not sure how that would work.
", "time": "2022-07-29T15:24:17Z"}, {"author": "Carsten Bormann", "text": "(CCast is a form of BIER for RPL that doesn't need bit number assignment.)
", "time": "2022-07-29T15:24:46Z"}, {"author": "Juliusz Chroboczek", "text": "Carsten, do you have a link?
", "time": "2022-07-29T15:25:03Z"}, {"author": "Juliusz Chroboczek", "text": "Lorenzo, yes.
", "time": "2022-07-29T15:25:32Z"}, {"author": "Carsten Bormann", "text": "https://datatracker.ietf.org/doc/html/draft-ietf-roll-ccast-01
", "time": "2022-07-29T15:26:59Z"}, {"author": "Juliusz Chroboczek", "text": "Noted, thanks.
", "time": "2022-07-29T15:27:09Z"}, {"author": "Carsten Bormann", "text": "(You were present when we presented that...)
", "time": "2022-07-29T15:27:13Z"}, {"author": "Juliusz Chroboczek", "text": "I must have missed the point, sorry.
", "time": "2022-07-29T15:27:32Z"}, {"author": "Zheng Zhang", "text": "To Juliusz: BIER can be used for virtual network interconnect, all the bits are set to make sure all the router can receive the packet, but the ingress router doesn't need to know the egress router in advance.
", "time": "2022-07-29T15:28:49Z"}, {"author": "Zheng Zhang", "text": "To Carsten, thanks for the link, I will read the draft.
", "time": "2022-07-29T15:29:25Z"}, {"author": "Carsten Bormann", "text": "CCast: a bit like BIER but without bit position assignment, provides true multicast based on Bloom Filters as a form of source routing
\nCcast is based on specific characteristics of RPL and can be very light-weight because of that
\nWork from 2013, essentially completed, but we lost our main implementation effort in 2018
\nThe pieces in draft-ietf-roll-ccast that weren't fully fleshed out are now essentially provided by draft-ietf-6lo-multicast-registration, so draft-ietf-roll-ccast could now go ahead easily
Zheng, but every router needs to know its own index, right? If two routers use the same index, packets will not reach the whole network, right?
", "time": "2022-07-29T15:29:41Z"}, {"author": "Juliusz Chroboczek", "text": "Carsten, right, I remember liking your presentation. (Bloom filters, yummy.)
", "time": "2022-07-29T15:30:19Z"}, {"author": "Zheng Zhang", "text": "Juliuze, yes, we can limit the receiver set through some configuration or policy
", "time": "2022-07-29T15:31:13Z"}, {"author": "Carsten Bormann", "text": "Filtering destinations requires storing mode, Pascal.
", "time": "2022-07-29T15:32:03Z"}, {"author": "Carsten Bormann", "text": "What stalled Ccast was trying to somehow generalize it, losing its actual advantages.
", "time": "2022-07-29T15:33:12Z"}, {"author": "Carsten Bormann", "text": "You need a term for evil duplicate suppression
", "time": "2022-07-29T15:51:18Z"}, {"author": "Carsten Bormann", "text": "The real duplicate suppression is based on sequence numbers added at ingress
", "time": "2022-07-29T15:51:50Z"}, {"author": "Henning Rogge", "text": "yes, its more \"hash based (no overhead) duplicate suppression\" vs. \"sequence number/source tuple\" duplicate supression
", "time": "2022-07-29T15:53:32Z"}, {"author": "Carsten Bormann", "text": "6loRH can minimize encaps overhead
", "time": "2022-07-29T15:53:34Z"}, {"author": "Juliusz Chroboczek", "text": "Henning: BIER has no seqno.
", "time": "2022-07-29T15:53:57Z"}, {"author": "Carsten Bormann", "text": "Yes, but the content-based duplicate detection needs the word \"evil\" in the name
", "time": "2022-07-29T15:54:02Z"}, {"author": "Carsten Bormann", "text": "Juliusz: because it encapsulates
", "time": "2022-07-29T15:54:23Z"}, {"author": "Juliusz Chroboczek", "text": "Agreed.
", "time": "2022-07-29T15:54:36Z"}, {"author": "Zheng Zhang", "text": "There is no duplication in BIER
", "time": "2022-07-29T15:54:42Z"}, {"author": "Henning Rogge", "text": "BIER is more like explicit multicast... it enumerate all destinations (in form of a bit-vector), right?
", "time": "2022-07-29T15:55:31Z"}, {"author": "Zheng Zhang", "text": "Yes
", "time": "2022-07-29T15:56:31Z"}, {"author": "Zheng Zhang", "text": "I mean that the BIER makes sure the packet be sent in the right out interface
", "time": "2022-07-29T15:57:58Z"}]