NTP WG Meeting Minutes

0900-1000, 13 July 2006

Karen O’Donoghue and Brian Haberman, co-chairs of the NTP WG, called the meeting to order. Dave Plonka volunteered to take minutes, and Yakov Stein agreed to act as jabber scribe. Karen addressed basic administrative issues including blue sheets and agenda bashing. The NTP WG deliverables and milestones were reviewed. All are currently behind schedule. The MIB has been published and is awaiting comment. The protocol and algorithms drafts have significant issues and will be the primary focus of this meeting.

Two key working group products have significant issues, the protocol document (draft-ietf-ntp-ntpv4-proto-02) and the algorithms document (draft-ietf-ntp-ntpv4-algorithms-01 (expired)). To facilitate in the production of these working group products, Brian and Karen visited Dave Mill’s in early July. At that meeting, they received a new NTP Reference and Implementation Guide document from Dr. Mills. Dr. Mills feels that this is a better representation of the current state of NTPv4 than the documents the working group had previously been working with. The missing pieces, according to Dr. Mills, are IPv6 and SNTP. Dr. Mills agreed to produce an update to the document including SNTP. Additionally, this document is not encumbered by the copyright concerns associated with the material received last year from Dr. Mills.

Brian and Karen proposed a way forward for these two key working group deliverables. This is shown in the meeting agenda slides. In summary, the plan is to:

• create an editor’s team,
• use the June 2006 Mill’s document as the starting point,
• add IPv6 and SNTP into that document, and
• produce an initial draft (or drafts) by September 1, 2006.

A number of questions related to this plan were discussed:

1.  Should the protocol and algorithms drafts be one document or two separate documents? 

In an ideal world, they should be separate documents. Separating the algorithms from the protocol will facilitate future development and will be easier to get through the IESG review process. However, the effort and delay required to create two separate documents may not be acceptable at this time. It was decided to task the editing group to evaluate, and implement if possible, the creation of separate protocol and algorithms documents from this baseline text.

2.  Should the resulting document or documents be Standards Track or Informational?

The protocol specification is definitely standards track. The group seemed divided on whether the algorithms document should be Informational or Standards track. Mark Townsley pointed out that the algorithms document could be Informational if it were not used as a normative reference in the protocol document. However, some felt that the algorithms are a key component and must be Standards track especially because they are integral to the NTPv4 implementation today. Others felt that in order to establish compliance one would need to reverse engineer the complex algorithms. Additionally, standardizing the algorithms would overstate the value of the current algorithm set. It was decided to task the editing team with evaluating the possibility of separating the documents (on which there is consensus) before the working group finalized whether the result should be Informational or Standards Track. 

The next agenda topic was the status of the possible IETF “PDF Experiment”. The NTP WG has volunteered to be one of the candidates to participate in this experiment if it gets approved in a timely manner. Brian has forwarded the NTP document received from Dr. Mills to the IESG so they can judge the complexity of rendering that information in ASCII. Mark Townsley and Stuart Bryant pointed out that this is a perfect example as RFC 1305 is the only existing non-ASCII RFC today.  There was discussion on clarification of the rules and some of the options for the NTP documents using these rules. In particular, it would be possible to make the figures external references and use psedo-code for the equations, but both these options degrade the readability of the document. Under the experiment, the PDF version could be normative.

To summarize the results of the meeting, the chairs indicated that there was consensus to establish an editing group using the June 2006 input from Dr. Mills as a starting point. The editing group is to consider the production of one document versus two and the possible IESG “PDF Experiment” during their efforts, but their first priority is to get drafts out in a timely manner.

The meeting was adjourned.