[{"author": "Anthony Somerset", "text": "

try refresh your browser and try again - that worked for Bob

", "time": "2023-03-29T00:31:04Z"}, {"author": "Anthony Somerset", "text": "

can cam be moved to the chairs?

", "time": "2023-03-29T00:37:16Z"}, {"author": "Anthony Somerset", "text": "

that was fast :joy:

", "time": "2023-03-29T00:37:33Z"}, {"author": "Naoki Matsuhira", "text": "

No Slide now?

", "time": "2023-03-29T00:37:48Z"}, {"author": "Eduard V", "text": "

slide is fine

", "time": "2023-03-29T00:38:00Z"}, {"author": "Behcet Sarikaya", "text": "

Moved to Chrommee, now good

", "time": "2023-03-29T00:38:33Z"}, {"author": "Naoki Matsuhira", "text": "

Solved. Thanks.

", "time": "2023-03-29T00:38:39Z"}, {"author": "Anthony Somerset", "text": "

there is definitely wierdness with the meetecho client going on - slides just dropped for me with an attention unstable connection popup
\ni am on the ietf wifi in the room

", "time": "2023-03-29T00:48:52Z"}, {"author": "Warren Kumari", "text": "

We can hear you!

", "time": "2023-03-29T00:49:02Z"}, {"author": "Tom Herbert", "text": "

Yes

", "time": "2023-03-29T00:49:17Z"}, {"author": "George Michaelson", "text": "

\"tommy can you hear me\" is a quote from Quadrophenia. An IPv6 address is 4x (a quad) of a v4. Coincidence? you decide

", "time": "2023-03-29T00:49:47Z"}, {"author": "George Michaelson", "text": "

(its not from quadrophenia, obviously its from tommy..)

", "time": "2023-03-29T00:50:07Z"}, {"author": "Warren Kumari", "text": "

Thank you adrian!

", "time": "2023-03-29T00:50:48Z"}, {"author": "Anthony Somerset", "text": "

we hear you Tom

", "time": "2023-03-29T00:51:07Z"}, {"author": "Warren Kumari", "text": "

Gee, thanks George, now I have that stuck in my head....

", "time": "2023-03-29T00:51:38Z"}, {"author": "Louis Chan", "text": "

For the VTN draft, since there is a \u201cS\u201d bit in the extension header, it could take more cycles to prcocess the packet in the forwarding plane. An evaluation of impact should be carried out.

", "time": "2023-03-29T01:07:46Z"}, {"author": "Behcet Sarikaya", "text": "

agree with Lorenzo

", "time": "2023-03-29T01:22:57Z"}, {"author": "Behcet Sarikaya", "text": "

small is beautiful

", "time": "2023-03-29T01:23:23Z"}, {"author": "Michael Richardson", "text": "

I heard Tom.

", "time": "2023-03-29T01:24:40Z"}, {"author": "Jen Linkova", "text": "

I suggest we discuss it in the WGLC

", "time": "2023-03-29T01:25:01Z"}, {"author": "Behcet Sarikaya", "text": "

Agreed Jen

", "time": "2023-03-29T01:25:39Z"}, {"author": "Lorenzo Colitti", "text": "

this is not about MPMH

", "time": "2023-03-29T01:25:44Z"}, {"author": "Lorenzo Colitti", "text": "

I think the algorithm is too complicated even in the simple case

", "time": "2023-03-29T01:26:00Z"}, {"author": "Behcet Sarikaya", "text": "

who is speaking? Name is not shown?

", "time": "2023-03-29T01:36:46Z"}, {"author": "Gorry Fairhurst", "text": "

Ana Custura speaking at mic

", "time": "2023-03-29T01:37:10Z"}, {"author": "Behcet Sarikaya", "text": "

Thanks Gorry

", "time": "2023-03-29T01:37:35Z"}, {"author": "Tom Herbert", "text": "

Thanksfor the comments, please chime in on the list!

", "time": "2023-03-29T01:41:58Z"}, {"author": "Shuping Peng", "text": "

@Behcet, @Gorry, name is added.

", "time": "2023-03-29T01:42:59Z"}, {"author": "Tom Herbert", "text": "

It's best to send all packets of a flow with consistent set of options if possibe, homogenuity is a benefit to minimizing drops or OOO

", "time": "2023-03-29T01:44:48Z"}, {"author": "Behcet Sarikaya", "text": "

Thanks Shuping

", "time": "2023-03-29T01:47:18Z"}, {"author": "Martin Horneffer", "text": "

the value probably varies a lot with exact HW at each hop.

", "time": "2023-03-29T01:49:40Z"}, {"author": "Anthony Somerset", "text": "

@Fernando Gont can you tune your mic gain down a bit, causing lots of clipping here

", "time": "2023-03-29T01:51:31Z"}, {"author": "Behcet Sarikaya", "text": "

+1 to Anthony

", "time": "2023-03-29T01:52:04Z"}, {"author": "Tom Herbert", "text": "

@Fernando, limits are in the eh-limits draft, processing recommendations are more in the hbh-processing. They are complementary for the most part

", "time": "2023-03-29T01:53:06Z"}, {"author": "Michael Richardson", "text": "

we want them to ignore the HbH, not igore/drop the packet!

", "time": "2023-03-29T01:54:59Z"}, {"author": "Fernando Gont", "text": "

IPv6 support by cloud providers is... say... tricky, at least :-)

", "time": "2023-03-29T01:55:00Z"}, {"author": "Michael Richardson", "text": "

them=hosts.

", "time": "2023-03-29T01:55:07Z"}, {"author": "Tom Herbert", "text": "

HBH-limits is clear that hosts must process all HBH and DestOPts

", "time": "2023-03-29T01:55:24Z"}, {"author": "Michael Richardson", "text": "

\"forward\" means, for hosts, pass it up to the next layer.

", "time": "2023-03-29T01:56:05Z"}, {"author": "Tom Herbert", "text": "

@Gorry, you might want to look at the Host/Intermediate Destinations/Intermeidate nodes in HBH-limits

", "time": "2023-03-29T01:56:15Z"}, {"author": "Erik Kline", "text": "

@Tom Herbert @Jen Linkova comment from an RTG AD about EH limits is that the document status might want to be BCP (I think it's currently PS).

", "time": "2023-03-29T01:57:26Z"}, {"author": "Jen Linkova", "text": "

Thanks Erik.

", "time": "2023-03-29T01:58:23Z"}, {"author": "Tom Herbert", "text": "

@Erid, yes we had that discussion on the list. The risk of BCP is that the draft would be less impactful and forceful

", "time": "2023-03-29T01:58:49Z"}, {"author": "Bob Hinden", "text": "

Meetecho: No speaker video

", "time": "2023-03-29T01:59:44Z"}, {"author": "Behcet Sarikaya", "text": "

Proposed Standard is OK, I think

", "time": "2023-03-29T01:59:51Z"}, {"author": "Anthony Somerset", "text": "

need to move the cam

", "time": "2023-03-29T02:00:10Z"}, {"author": "Bob Hinden", "text": "

Meetecho: Thanks!

", "time": "2023-03-29T02:00:31Z"}, {"author": "Lorenzo Miniero", "text": "

Apologies, missed the speaker change, just fixed

", "time": "2023-03-29T02:00:35Z"}, {"author": "Behcet Sarikaya", "text": "

@Bob, I think Shuping is remote

", "time": "2023-03-29T02:01:08Z"}, {"author": "Behcet Sarikaya", "text": "

Now we see her

", "time": "2023-03-29T02:01:40Z"}, {"author": "Anthony Somerset", "text": "

shuping in the room :)

", "time": "2023-03-29T02:01:54Z"}, {"author": "Behcet Sarikaya", "text": "

I thought I saw Shuping on site at Dispatch WG meeting

", "time": "2023-03-29T02:03:05Z"}, {"author": "Behcet Sarikaya", "text": "

she registered as remote?

", "time": "2023-03-29T02:03:43Z"}, {"author": "Behcet Sarikaya", "text": "

Too late for me here so I have to go

", "time": "2023-03-29T02:05:53Z"}, {"author": "Shuping Peng", "text": "

hi Behcet, I am on site finally after a few years, :D

", "time": "2023-03-29T02:08:51Z"}, {"author": "Bob Hinden", "text": "

Behcet: Shuping was here in person :-)

", "time": "2023-03-29T02:08:55Z"}, {"author": "Anthony Somerset", "text": "

i saw a related question earlier in the week in the related presentation

\n

how will routing between hosts on same network work when doing /64 per host PDs?

", "time": "2023-03-29T02:11:16Z"}, {"author": "Ole Tr\u00f8an", "text": "

Anthony: Would have to be via the default router.

", "time": "2023-03-29T02:11:53Z"}, {"author": "Jen Linkova", "text": "

and the router may - or may not - send redirects. Many networks blocks peer2peer anyway

", "time": "2023-03-29T02:12:20Z"}, {"author": "Ole Tr\u00f8an", "text": "

Seems like simple to attack, if a host is misbehaving, or even just forgetting to release the assigned prefix back.

", "time": "2023-03-29T02:12:45Z"}, {"author": "Michael Richardson", "text": "

but, if the network does like peer2peer (DC use with VMs), then how do we optimize the routes?

", "time": "2023-03-29T02:13:05Z"}, {"author": "Jen Linkova", "text": "

Well, you might want your hosts to support RIOs then ;)

", "time": "2023-03-29T02:13:25Z"}, {"author": "Ole Tr\u00f8an", "text": "

Michael: If it sounds like a duck... it might be a router and it could run a routing protocol. We just need the new 'OSPF' flag in the RA. ;-)

", "time": "2023-03-29T02:14:15Z"}, {"author": "Michael Richardson", "text": "

Ole, yes.

", "time": "2023-03-29T02:14:39Z"}, {"author": "Michael Richardson", "text": "

I don't think we must pick one.

", "time": "2023-03-29T02:16:15Z"}, {"author": "Michael Richardson", "text": "

Pick At Least One.

", "time": "2023-03-29T02:16:21Z"}, {"author": "Nick Buraglio", "text": "

For scaling issues I can see this having several advantages around resource scaling in both filtering and route table. This solves a few uses cases I have in the DC and development environment as well.

", "time": "2023-03-29T02:16:35Z"}, {"author": "Sri Gundavelli", "text": "

What is the privacy implication of using a per-device prefix in enterprise environments?

", "time": "2023-03-29T02:16:38Z"}, {"author": "Ole Tr\u00f8an", "text": "

Well, then it will fail when it's connected to a link with two routers on the same link. This is the problem of \"link\" ownership in prefix assignment.

", "time": "2023-03-29T02:17:00Z"}, {"author": "Tom Hill", "text": "

/me braces for Fernando's mic

", "time": "2023-03-29T02:17:15Z"}, {"author": "Tom Hill", "text": "

Oh that's a bit better

", "time": "2023-03-29T02:17:32Z"}, {"author": "Tom Hill", "text": "

Less distorted

", "time": "2023-03-29T02:17:42Z"}, {"author": "Anthony Somerset", "text": "

thanks @Fernando Gont

", "time": "2023-03-29T02:18:00Z"}, {"author": "Nick Buraglio", "text": "

Ole Tr\u00f8an said:

\n
\n

Michael: If it sounds like a duck... it might be a router and it could run a routing protocol. We just need the new 'OSPF' flag in the RA. ;-)

\n
\n

This is how we labbed up some of the DC issues, but it has the oppositie sc
\naling issue - it requires far too much from an operational perspective, and has some \"soft ceiling\" issues with running routing daemons on hosts for some folks.

", "time": "2023-03-29T02:18:21Z"}, {"author": "Anthony Somerset", "text": "

i'm only seeing good benefits in enterprise LAN
\nlimited benefit in DC space really especially the reverse scalingg effect

", "time": "2023-03-29T02:22:16Z"}, {"author": "Nick Buraglio", "text": "

For the DC it can simplify a containerized environment, but I agree the real benefits are in the enterprise / campus realm.

", "time": "2023-03-29T02:23:40Z"}, {"author": "Jen Linkova", "text": "

actually in DC - if you are using EVPN/VXLAN - that's very useful

", "time": "2023-03-29T02:24:27Z"}, {"author": "Jen Linkova", "text": "

but I'm obviously biased ;)

", "time": "2023-03-29T02:25:28Z"}, {"author": "Dave Thaler", "text": "

Is this presentation about https://xkcd.com/927/ ?

", "time": "2023-03-29T02:27:11Z"}, {"author": "Nick Buraglio", "text": "

that's a good point

", "time": "2023-03-29T02:27:32Z"}, {"author": "Greg Mirsky", "text": "

I wonder why is it Geneve is missing from the analysis?

", "time": "2023-03-29T02:27:36Z"}, {"author": "Greg Mirsky", "text": "

Strike it, too early ;)

", "time": "2023-03-29T02:28:20Z"}, {"author": "Nick Buraglio", "text": "

The PD option also solves a non-trivial issue I see a lot, too, which is a desire for IPv6-only networks to leverage DHCPv6 but feel that they cannot due to lack of support in android, this would give them a path to their goals. I hear this complaint several times a month in the US migration to IPv6-only projects.

", "time": "2023-03-29T02:28:39Z"}, {"author": "Nick Buraglio", "text": "

that's a good point

", "time": "2023-03-29T02:29:00Z"}, {"author": "Philipp Tiesel", "text": "

AWS already uses PD to get a /80 to instances. It is really useful to bootstrap prefixes for containers

", "time": "2023-03-29T02:29:37Z"}]