[HN Gopher] Unicast Use of the Formerly Reserved 127/8
___________________________________________________________________
Unicast Use of the Formerly Reserved 127/8
Author : todsacerdoti
Score : 12 points
Date : 2021-11-16 21:13 UTC (1 hours ago)
(HTM) web link (www.ietf.org)
(TXT) w3m dump (www.ietf.org)
| dtaht wrote:
| I note the 240/4 and 0/8 and especially zeroth ideas are the more
| viable of this bunch.
|
| What I mentally see as a possible use for 127/8 (with 127/16
| still held in reserve) - is the really painfully untransparent
| string of vm -> container -> OS -> offload engine that we have
| little insight into today.
| rurcliped wrote:
| It's common for outbound traffic to 127/8 to be blocked, at the
| application layer, to prevent SSRF. For example, in WordPress:
|
| https://github.com/WordPress/WordPress/blob/9d4d3f8fba13be09...
| op00to wrote:
| There are so many embedded systems that assume so much about
| 127/8. I don't understand what industry is driving this that
| can't simply look to ipv6...
| toast0 wrote:
| Contrary to probably most people, I've actually got experience
| using a lot more than 127.0.0.1. I needed to support an HTTPS
| service built as a TLS terminating proxy plus an HTTP server on
| the same machine, the HTTP server didn't understand unix sockets,
| and the required max connections goal was 1 million; it would
| have probably been faster and more effective to teach the http
| server how to do unix sockets or how to do TLS efficiently, but
| there was organizational opposition to both at the time. Of
| course, both are now options.
|
| So that means lots of connections on localhost, but you run into
| difficulties with managing ports unless you listen on multiple
| ips. Strategically picking the IPs to listen to helps reduce lock
| contention in the tcp stack as well. IIRC, I think we used
| 127.0.X.1 (which would be compatible with this proposal), but I
| no longer have access to the specifics.
|
| The single loopback address of IPv6 is kind of insufficient, but
| there's not a whole lot of need for IPv6 over localhost, so if I
| ran into limits, I'd just use IPv4, or even use one of the IPv4
| in IPv6 space mappings with IPv4 loopback addresses.
|
| OTOH, if they're looking at opening up 127/8, why not also open
| up 0/8 (other than 0/24) and the class E 240/4 (other than
| 255.255.255.0/24)? There's some use of class E addresses[1], but
| I don't think it's that much more of a disruption than 127/8.
|
| [1] Cloudflare uses them as Pseudo addresses for IPv6 transition,
| and I think I've seen that elsewhere too.
| https://blog.cloudflare.com/supporting-the-transition-to-ipv...
| schoen wrote:
| > OTOH, if they're looking at opening up 127/8, why not also
| open up 0/8 (other than 0/24) and the class E 240/4 (other than
| 255.255.255.0/24)? There's some use of class E addresses, but I
| don't think it's that much more of a disruption than 127/8.
|
| That's exactly what we're proposing! This draft is just one of
| four drafts. The other three are
|
| https://datatracker.ietf.org/doc/draft-schoen-intarea-unicas...
|
| https://datatracker.ietf.org/doc/draft-schoen-intarea-unicas...
|
| https://datatracker.ietf.org/doc/draft-schoen-intarea-unicas...
| schoen wrote:
| Hey, I'm the lead author of this draft!
|
| As I mentioned below, this is just one of four proposals that
| we're making at IETF about reclaiming unused IPv4 address space.
| The other three proposals are
|
| https://datatracker.ietf.org/doc/draft-schoen-intarea-unicas...
|
| https://datatracker.ietf.org/doc/draft-schoen-intarea-unicas...
|
| https://datatracker.ietf.org/doc/draft-schoen-intarea-unicas...
|
| In each case, the address space in question was reserved a long
| time ago (during the 1980s) for reasons that have not proven
| necessary or that are no longer applicable.
|
| It is indeed a big undertaking to change this behavior
| everywhere, but we propose that it will be a gradual process. It
| turns out that 240/4 is already widely supported (implementers
| liked the idea when it was previously proposed at IETF in 2008),
| while we've also gotten patches into Linux for the 0/8 and lowest
| address behavior, and now in FreeBSD for the lowest address
| behavior as well. (The lowest address was originally reserved for
| an "alternative broadcast address" because 4.2BSD chose it as the
| segment-directed broadcast address in 1983, before there was a
| standard to say which address to use for this purpose.)
|
| There are lots of other complexities and history that I'm happy
| to talk about if people are interested.
|
| The amount of IPv4 space still "reserved for future use" in 240/4
| is 228 addresses, or 1/16 of all of IPv4!
|
| Edit: people who are especially interested in this topic can also
| watch me presenting on this for 15 minutes at IETF112 last week.
|
| https://youtu.be/cqPVdBvgiXI?t=409
| op00to wrote:
| Nope.
___________________________________________________________________
(page generated 2021-11-16 23:01 UTC)