Presentation is loading. Please wait.

Presentation is loading. Please wait.

ENUM Implementation Issues Lawrence Conroy

Similar presentations


Presentation on theme: "ENUM Implementation Issues Lawrence Conroy"— Presentation transcript:

1 ENUM Implementation Issues Lawrence Conroy Lwc@roke.co.uk

2 11-Nov-20031 Things we have seen… There are a number of different NAPTRs “out there” Not all of them are compliant to RFC2916bis It’s an implementation issue how one deals with these But…one should be liberal with what one accepts There’s a list of 8 issues we hit whilst building and testing an RFC2916bis compliant client (and Admin system) - they’re on the ENUM ML Proposal: “living” document to capture the issues and how folks deal with any ambiguities - this is a kind of BCP/Implementor’s Guide - NOT a protocol standard

3 11-Nov-20032 For the email-challenged (a) Issue 1 - Case Sensitivity –Are ENUMservices case sensitive or not? –Are flags case sensitive or not? [We treat them both as case insensitive] Issue 2 - E2U at which end? –Some folk publish RFC2916 NAPTRs, some use 2916bis; this means that E2U could be at either end of the service field –One “quick fix” is to treat E2U as a reserved string, so it is NOT a valid ENUMservice [We process everything as an ENUMservice, and then strip the E2U “pseudo-service”]

4 11-Nov-20033 For the email-challenged (b) Issue 3 - Non-finals with or without empty service field? –Should the client check and use the service field in a non-final? –Must a registrant populate non-finals with an empty service field? [We publish them with empty field and ignore it in the client] Issue 4 - Does Order “count” across domains? –If one traverses a non-final reference to another domain, AND that domain has NAPTRs with a higher order value BUT there are none appropriate for the client, SHOULD one discard any unprocessed NAPTRs in the “referring” domain? [We treat order as significant only within a domain]

5 11-Nov-20034 For the email-challenged (c) Issue 5 - Non-final loop treatment –If there’s a loop in a chain of non-final references, then should one: Give up on the whole query Give up on the reference and carry on as if it wasn’t there? Give up on the whole “nested” set of references and return to any NAPTRs in the original domain? [We give up on the reference and carry on - (option 2)] Issue 6 - Non-final loop detection? –Does one need a full “domain traversed” list, or can on “get away” with a simple ‘recursion count’ for loop detection? –If the latter, what’s a reasonable limit for the number of non-final –references [We keep a count and give up on 5 nested references]

6 11-Nov-20035 For the email-challenged (d) Issue 7 - Treatment of NAPTRs with identical preferences –If there are several NAPTRs with the same preference, then must the client randomise its processing of the NAPTRs? [We process them in the order they arrive] Issue 8 - L2R or R2L processing of service field? –If a NAPTR has more than one ENUMservice in its service field, then should the client process them from the left to the right or the other way around? [We process them from left to right]


Download ppt "ENUM Implementation Issues Lawrence Conroy"

Similar presentations


Ads by Google