Operational Requirements Area Report Directors: o Scott Bradner o Mike O'Dell Four Operational Requirements Working Groups met during the Dallas IETF meeting and there were five BOF sessions held. BMWG AM Session, December 5, 1995 The BMWG met Tuesday morning to work on the device benchmarking agenda. Jim McQuaid led the session. The group discussed the status of the current I-D on network element benchmarking. The recent changes and several small open issues were discussed and resolved. The group feeling is to complete the current I-D as an informational RFC and to let new methodology documents add methods or definitions as needed in the future, rather than revise and supersede this document constantly. Initial discussions were held on Ethernet switch testing methods and the benchmarking of call setup in SVC networks. BMWG PM Session, Dec 5, 1995 The BMWG met late Tuesday afternoon to work on the IPPM agenda. Guy Almes led the session. After a brief review of history and status, discussion focused on four areas: o measures of delay, o measures of flow capacity, o (very brief) measures of availability, and, o the use of probe/transponder machines. The formal session concluded with a talk by Steve Corbato. CIDRD Working Group The CIDRD Working Group met at the Dallas on Wednesday, December 6. Tony Li gave a summary of the IPv4 address space allocations, with extrapolations giving an Ipv4 address space lifetime of 2018 +/- 8 years. Fred Solensky presented a logistical analysis of the usage of the 128/2 and 192/3 address space projecting that 128/2 usage will stabilize at 62% of the total and that 192/3 usage will stabilize at 90% by 2006. Questions were raised whether this sort of projection is meaningful. Bill Manning reported that 13% of the 0/1 space has been reclaimed during the past six months. Erik-Jan reported on the size of the global routing table, which has grown approximately 10% in the past four months. The routing table represents over 822,000,000 potential addresses of which approximately 8,000,000 are believed to be active. This represents a utilization efficiency of less than 1%. The status of a number of CIDRD documents was considered and debated: Appeal to Return Unused Address Space (submit as BCP), RFC1597bis (final revisions, submit as BCP), Address Ownership (revise, further review), Class A Subnet Deployment Considerations (submit as Informational), Net 39 Experiment Report (revise, review, submit as Informational), CNAME Extensions to IN-ADDR (rehome to DNSIND Working Group). Much discussion of Non-Local Aggregation occurred, Internet-Draft to be crafted. Yakov Rekhter presented on the subject of Charging for Routing Advertsements. Much discussion followed the presentation. PIER BOF The PIER BOF was held in two sessions. The first covered the draft charter, which was updated to reflect a broader scope of working group interaction with other working groups and a clarification on the focus on IPv4. A review of a series of documents that had already been discussed was concluded. It was noted that most of the documents reflected cookbook approaches to renumbering small sites. A wide ranging discussion on renumbering ensued Participants dicussed how it applies to Routing & Infrastructure, Firewalls, Multihomed sites, Servers, API's, and Tools. A number of participants indicated that they were either currently involved in a renumbering project or were going to start one very soon. There was agreement that a working group (or two) ought to be formed to address some of these issues. The second session was used to generate milestones. Six classes of documents were identified: o How to renumber small sites; o A Goal statement; o Where IP addresses are used in applications; o Tool catalog; o Case studies; and, o How renumbering affects my network presence. Volunteers were "persuaded" to tackle at least one item for each group. It is expected that a majority of the work will be completed by Summer 1996. RADIUS BOF The RADIUS BOF had 102 attendees. Suggestions were made to have a NAS-Port-Type instead of the recently proposed NAS-Port-Id. This will be followed up on the mailing list. A few clarifications were suggested for Draft 01 of RADIUS and RADIUS accounting. The attendees were very supportive of the idea of attempting to move the RADIUS draft to Proposed Standard in January, and if that wasn't possible in a relative timeframe, then to issue it as an Informational RFC in January. Then, a continuing effort would be made on getting it placed on a Standards Track. Members were very supportive of the idea of amending the charter to start work on a separate Internet-Draft concerning RADIUS extensions, as long as that wouldn't imperil the standards-track status of the current RADIUS draft. The Working Group charter will be revised accordingly. Interest was also expressed in holding a RADIUS Bake-Off in the Spring. RPS Working Group A mechanism to support Destination Preference Attribute was presented, along with the pros and cons of using AS-Macros or AS- Expressions in specifying AS Peerings in policy terms. Curtis Villamizar presented what needs to be specified to handle aggregation and different alternatives to achieve it. There was no consensus at the meeting, but consensus should be reached on the mailing list. There was a debate on the mailing list whether the rpsl as path regular expressions should be character (like Cisco) or integer (like gated) based. The rpsl syntax was presented and compared to both Cisco and gated syntaxes. Several ISP representatives who actually use Cisco routers suggested to stick with integer based expressions. There was a unanimous agreement. Deborah Estrin's work on multi-cast policies was presented. Tony Bates suggested that it be published as an informational RFC. Extensions needed to support and take full advantage of SDRP were suggested. David Kessens' work on Real Time Database mirroring was presented. Jerry Scharf presented his intentions to implement a database format independent transport mechanism. RTFM BOF Fifty seven people attended the BOF. Four presenters covered the following topics: the need for flow measurement from ISP and large customer perspectives; and, existing implementations of the Meter MIB experiences with the existing measurement tools. Consensus was that a standard way of collecting flow data is highly desirable and that a traffic measurement working group be formed to review existing proposals and produce a standards track document set. A discussion list will be established: . In addition, draft charter was prepared. RWhois WG Scott Williamson presented the working group with a summary of the current status of the implementation. The working group then discussed what is required in making RWhois operational based on the current experiences and to develop planned steps to make it robust. The Working Group then discussed changes to the initial version of the protocol and ways of entering authentication within rwhois for updates and new records. Then followed a discussion about reducing flat spaces using DNS. The Working Group concluded with a call for volunteers to help develop RWhois. STDGUIDE BOF The draft of the guide was presented at the meeting and discussed. A new version incorporating input from the meeting will be prepared by the authors and sent out on the mailing list.