This archive is retained to ensure existing URLs remain functional. It will not contain any emails sent to this mailing list after July 1, 2024. For all messages, including those sent before and after this date, please visit the new location of the archive at https://mailman.ripe.net/archives/list/[email protected]/
[dns-wg] [dns-operations] Announcement - DNS flag day on 2019-02-01
- Previous message (by thread): [dns-wg] [dns-operations] Announcement - DNS flag day on 2019-02-01
- Next message (by thread): [dns-wg] [dns-operations] Announcement - DNS flag day on 2019-02-01
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Florian Weimer
fw at deneb.enyo.de
Thu Jun 14 18:54:24 CEST 2018
* Paul Vixie: > Florian Weimer wrote: > ... >> >> Or you can avoid fragmentation in the first place, ... > > so, just always use tcp if you're expecting more than 1200 octets? Sure, unless it's a special dedicated network etc. >> Theoretically, even with a 1200-byte EDNS buffer size, ... > > but, there are useful answers larger than that. A lot of those answers are artificially inflated with data that clients cannot use anyway. >> Another benefit of this change is that many of the ENDS-related >> problems go away. > > i'd rather we broke everything that won't let edns through. the internet > can't grow in avoidance-mode. we have to confront, too. It's not clear to moe at all what value larger packets provide. You may avoid the occasional TCP transaction. But then there are weird corner cases, such as the difficulty of serving UDP over IPv6 in a stateless fashion, while still supporting fragmentation.
- Previous message (by thread): [dns-wg] [dns-operations] Announcement - DNS flag day on 2019-02-01
- Next message (by thread): [dns-wg] [dns-operations] Announcement - DNS flag day on 2019-02-01
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
[ dns-wg Archives ]