Another potential application/use-case for this is cloud provider (eg, VM hosting) networks which often provide only an L3 abstraction for default networking. All of them seem to be coming up with different solutions in this space.
", "time": "2023-07-25T16:59:37Z"}, {"author": "Erik Nygren", "text": "(this == non-broadcast address, which is what many/most cloud hosting providers have for their publicly-addressed networking)
", "time": "2023-07-25T17:01:24Z"}, {"author": "\u00c9ric Vyncke", "text": "I misread WLCG into WGLC...
", "time": "2023-07-25T17:24:35Z"}, {"author": "Michael Richardson", "text": "Sounds like a BCP? Or just Informational on \"we did X\"
", "time": "2023-07-25T17:30:54Z"}, {"author": "\u00c9ric Vyncke", "text": "'informational' should be the target
", "time": "2023-07-25T17:31:14Z"}, {"author": "\u00c9ric Vyncke", "text": "(I think)
", "time": "2023-07-25T17:31:21Z"}, {"author": "\u00c9ric Vyncke", "text": "@Lorenzo, some routers DO look into parts of the flow label
", "time": "2023-07-25T17:31:50Z"}, {"author": "Cheng Li", "text": "Impressive work.
", "time": "2023-07-25T17:53:19Z"}, {"author": "Michael Richardson", "text": "I'd like to kick off a series of design team meetings to actually detail these objections. This is a really serious problem that we need to address.
", "time": "2023-07-25T17:58:27Z"}, {"author": "Michael Richardson", "text": "(on IPv6 Site connection to many Carriers )
", "time": "2023-07-25T17:58:40Z"}, {"author": "Cheng Li", "text": "agree, serious problem to be addressed
", "time": "2023-07-25T17:59:10Z"}, {"author": "Erik Nygren", "text": "Agreed that this an important problem to solve and have some good recommendations around.
", "time": "2023-07-25T17:59:12Z"}, {"author": "Erik Nygren", "text": "One thing we might want to consider is treating active/active cases as separate from primary/failover use-cases. There might be solutions that work for primary/failover where failover has degraded functionality. (eg, and where primary has no NAT but failover does have NAT and may have some things break until the primary comes back)
", "time": "2023-07-25T18:01:13Z"}, {"author": "Michael Richardson", "text": "Michael Richardson said:
\n\n\nI'd like to kick off a series of design team meetings to actually detail these objections. This is a really serious problem that we need to address.
\n
(I still cry about SHIM6)
", "time": "2023-07-25T18:01:16Z"}, {"author": "Eduard V", "text": "If MHMP problem so complex, then document is needed even more.
", "time": "2023-07-25T18:02:41Z"}, {"author": "Paolo Nero", "text": "Re: policy table for ULA issue, this is mentioned in the draft
", "time": "2023-07-25T18:03:26Z"}, {"author": "Michael Richardson", "text": "Extension Headers suffer the same problem as Multihoming. No solutions can get traction because there are no uses because nobody can get traction.
", "time": "2023-07-25T18:10:51Z"}, {"author": "Michael Richardson", "text": "I concur, we need some document that says something. I don't insist it be an RFC vs a wiki page. If someone could actually use the RFC for a procurement decision, then make it an RFC, but I don't get the impression we are at that place yet. I suspect that the underlying protocol that uses the EH is already useable for procurement, we just need to keep a collection of pointers.
", "time": "2023-07-25T18:13:13Z"}, {"author": "Geoff Huston", "text": "Bob is correct - HBH is not well supported, but DST has a better track record - see https://stats.labs.apnic.net/v6frag for a more comprehensive report on drop rates for Frag, DST and HBH drop rates in the Ipv6 Public Internet
", "time": "2023-07-25T18:15:19Z"}, {"author": "Tim Wicinski", "text": "Wait why is my name being called out?!>
", "time": "2023-07-25T18:16:50Z"}, {"author": "Tim Wicinski", "text": "wait, I didn't break anything (not a work meeting)
", "time": "2023-07-25T18:17:26Z"}, {"author": "Tim Wicinski", "text": "DNSOP also offered to hold an interim on this doc if neeeded.
", "time": "2023-07-25T18:23:07Z"}]