Download presentation
Presentation is loading. Please wait.
Published byEllen Flowers Modified over 8 years ago
1
2007/07/23IETF69 enum-combined1 draft-ietf-enum- combined IETF69 Otmar Lendl Michael Haberler Richard Stastny
2
2007/07/23IETF69 enum-combined2 Motivation The long-term solution (new apex for I-ENUM) may take some time to materialize. Interim solution to get an interoperable I-ENUM setup up and running.
3
2007/07/23IETF69 enum-combined3 Idea Initial idea: Leveraging the e164.arpa infrastructure by branching off the e164.arpa tree Q: Where to branch? Usually at the CC level Maybe at the NPA level
4
2007/07/23IETF69 enum-combined4 Evolution: -00: external table listing where to branch -01: DNS record at the CC level indicates where to branch (need an integer value, stored in ??) If not found, look down the tree Strong recommendation from DNS ppl: do your own RRTYPE Dallas treaty: Generic RRTYPE combined draft uses that facility
5
2007/07/23IETF69 enum-combined5 Prague Pre-IETF68 (-03): EBL definition + ENUM w/ EBLs DDDS application Combined draft: use-case of EBLs, EBL located at “infrastructur.CC.164.arpa”, walk down tree if not found. dnsext interaction: RFC 2929bis experiment Long chat with Ed Lewis and Olafur tree walking location of EBL
6
2007/07/23IETF69 enum-combined6 -05 no more “infrastructure.3.4.e164.arpa EBL..”. EBL at CC domain itself. -> Need a RRTYPE for each use-case of EBLs. There is no such thing as a “generic EBL”. Got rid of DNS tree walking. Good DNS karma Walking down was less a problem then walking up
7
2007/07/23IETF69 enum-combined7 AD-Review Jon: “please be more specific re: motivation” Ok, so let’s play devil’s advocate: The IEBL contains three fields: SEPARATOR POSITION APEX
8
2007/07/23IETF69 enum-combined8 SEPARATOR “What label to insert in order to branch” E.g. “i” leads to 6.1.4.6.5.0.5.1.i.3.4.e164.arpa Do we really need it? Is it a hard requirement that countries are able to use different branch label? Not really. Nice, but not a MUST.
9
2007/07/23IETF69 enum-combined9 POSITION “Where to insert the SEPARATOR” Usually the CC length, e.g. 2: XX.3.4.e164.arpa 3:XX.3.5.3.e164.arpa Might be different for group-of-country codes: 4:XX.N.P.A.1.e164.arpa But: in such cases the GoC have to agree on the EBL (remember: no more tree-walking!), so why not agree on double delegations like N.P.A.1.e164.arpa and N.P.A.XX.1.e164.arpa Thus: the EBL does not buy a NPA independence from the GoC any more. We already need the prior knowledge of the CC lengths, so why not just make that table be the POSITIONs?
10
2007/07/23IETF69 enum-combined10 APEX “what replaces e164.arpa” Facilitate the transition to the long-term solution. 3.4.e164.arpa IEBL 0 “” i-enum.arpa *. Not all countries will switch at the same moment. Older clients MUST be redirected to the new tree. That’s a known problem in the DNS space. Solution: XX.3.4.e164.arpa DNAME 3.4.i-enum.arpa * whatever the final apex will be.
11
2007/07/23IETF69 enum-combined11 So we’ve come full circle By removing some features, a simpler solution is possible: Always branch at a statically configured position, using the same label. Use DNAMEs for the transition to the long-term solution.
12
2007/07/23IETF69 enum-combined12 Hum 1 What should we do: Keep the EBL work, polish the drafts and off to the IESG we go. or Junk the EBL draft, rewrite the combined draft to use simpler solution. WG Last call ASAP.
13
2007/07/23IETF69 enum-combined13 Hum 2 (if EBLs are junked) What (fixed) label to use for branching: “i” “infrastructure” “ienum” “i-enum”
14
2007/07/23IETF69 enum-combined14 Hum 3 (if EBLs are junked) Fix the location on the CC length. vs. Define an exception for +1 (branch at NPA level).
15
2007/07/23IETF69 enum-combined15 Reserve Slide: What was the issue with DNAME? Can’t use DNAME if POSITION > CCLEN, e.g. i.1.e164.arpa DNAME XX.N.P.A.1.e164.arpa
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.