The Chair introduced the SIG and explained the agenda.
- Desired IRR operational model ?IRR/Whois interaction
Kuniaki Kondo, JPNIC
This presentation outlined the JPNIC IRR activities, including the JPNIC IRR Workshop. Recommendations were made as to the general requirements for establishing and operating an IRR.
Questions and discussion
No questions or discussions.
Top
- IRR pilot project status report
George Michaelson, APNIC
This presentation provided an update of APNIC's IRR activity since the previous Open Policy Meeting, including the deployment of a RIPE v3 test bed server combined with a MySQL service.
Questions and discussion
It was explained that there may need to be a procedure developed in relation to who is able to mirror the IRR. It was noted that the threshold for access control is likely to be on a /24 boundary.
Top
- New version of the RIPE database
Andrei Robachevsky, RIPE NCC
This presentation provided an update of the status of version 3 of the RIPE database, including statistics relating to use, development of software tools, and migration issues.
Questions and discussion
It was noted that this has been a very successful software release and is very easy to work with.
It was noted that MySQL does show some difficult behaviour in the implementation. It was explained that in theory, other relational databases, such as Oracle, should be possible to use.
It was noted that a huge amount of the load on the databases seems to be coming from automated services such as Zone Alarm and similar products. It was noted that this may not be a problem in the full deployment. It was also noted that there is the facility for denying access to extreme users.
There was a discussion of the authorisation procedures in relation to the creation of route objects. It was noted that PGP authorisation is the simplest method.
It was noted that support of IPv6 route objects is not yet formerly planned, but it has been included on the task list.
It was noted that there is work on an object library and that the dictionary may be compiled on one file. It was suggested that the dictionary is likely to be very stable.
Top
- Database privacy issue
Seung-Min Lee, I-Names
This presentation outlined a proposal relating to database privacy policy, developed by the Broadband Working Group in response to concerns relating to the privacy implications of registering every assignment, especially in cable and xDSL networks.
Questions and discussion
There was a comment strongly agreeing with the option of registering the ISP contact. It was noted that in Japan, most of the email to registered contacts is in English and would be likely to be ignored by the end-user.
It was explained that perhaps the easiest way to distinguish between residential and business users would be to consider the type of service contract.
It was also noted that there may be a need to also increase the level of responsibility that ISPs accept for the quality of registration information relating to their network assignments.
There was a show of hands on the proposal to amend section 7.17 of APNIC-076 to state that every network assignment greater than /30 must be registered; and that
assignments of /30 or smaller, including host assignments, may be registered at the discretion of the end-user and ISP. Consensus was reached to adopt this proposal.
There was a show of hands on the proposal to amend section 7.17.2 of APNIC-076 to state that in the case of residential users only, there should be the option to register the ISP's technical contact as the administrative contact (admin-c). Consensus was reached to adopt this proposal, but it was noted that there should be more discussion to consider extending the policy to non-residential users also.
Top
- Status of prov-reg and whoisfix BOF IETF discussions
Joao Damas, RIPE NCC
This presentation gave a brief overview of the Prov-reg Working Group and the Whoisfix BOF at the recent IETF in London.
Questions and discussion
There was a brief discussion relating to the relevance of Routing Registry data. It was noted that it is not likely that whois will be changed to allow cross-database services.
It was noted that the registry community does not wish to overload whois by pushing for additional unnecessary features.
Top
- Origin field
Xing Li, CERNET
This presentation described a proposal to add an “origin?field to inetnum objects and request forms.
Questions and discussion
It was noted that this proposal may help to increase the compliance with the ideal practice of showing a prefix coming from a single AS.
It was noted that the current practices relating to multihoming may complicate the implementation of this proposal.
It was noted that in practice, it is better to use the route object to do filtering. There was a discussion related to this issue. There was a suggestion that this information could be included in the description field instead and may be more useful due to the multihoming issues.
It was suggested that making this field mandatory for IPv6 may assist in deployment of IPv6.
The Chair recommended that this discussion should be taken up on the discussion list.