802.3z interim meeting minutes, Tuesday Sept. 30/Wednesday Oct. 1, 1997 Santa Clara, CA ------------------------------------------------------------------------ Tuesday Sept. 30, 1997 9:15 HF Chair, Howard Frazier, calls meeting to order, 9:15 AM. Howard starts the meeting by presenting the Agenda; slide 1/7, cover sheet agenda slide 2/7, item 1: Welcome and introductions agenda slide 2/7, item 2: Select recording secretary Tom Mathey from Baynetworks volunteers 9:38 HF agenda slide 2/7, item 3: P802.3z/D3.1 WG ballot update refer to agenda handout discussion on why it is important to reduce unresolved TR (technical required) comments 9:41 HF agenda slide 2/7, item 4: Standards development timeline refer to agenda handout Howard provides a review of schedule, no change to timeline targets followed by what is really important distinguish between opinion vs real problem or flaw we must have a functional standard importance of “last technical change” again, Howard emphasized the need for attention to detail, formal documentation, and procedures/process during the last steps. 9:44 HF agenda slide 2/7, item 5: Comment Resolution refer to agenda handout Howard emphasized the objective and direction of this meeting as: this is comment resolution time focus on resolving a few items do not thrash old issues need explicit detail want no surprises as we go forward 9:49 HF agenda slide 2/7, item 6: Distribution of documents Handouts #1. P802.3z/D3.1 Comment Resolution II #2. P802.3z/D3.1 hot items report #3. P802.3z/D3.1 report by comment ID #4. P802.3z/D3.1 Comments, 178 page on technical comments #5. P802.3z/D3.1 Comments, 154 page on editorial comments #6. Maui, unconfirmed minutes #7. London, unconfirmed minutes 9:49 HF agenda slide 2/7, item 7: “Big Ticket” items those that cross clauses or have major impact Howard then turns meeting over to Dr. Howard Johnson 9:49 HJ our chief technical editor gave a nice (5 slides) presentation on the process for “your comment” from submittal thru resolution. Howie discussed the submittal process, handling of comments, status codes, and available reports. Howard also invites suggestion, comments, or other on the data base. Refer to handout #1 titled: P802.3z/D3.1 Comment Resolution II. 9:59 done 10:04 document handout continues 10:08 HJ our chief technical editor provides a (3 slides) line by line short review of each of the 34 ajor ticket items which need to be covered by each sub-working group, covers which subclauses are included, and generally sets the direction he wants to subgroups to follow. The presentation and slides are covered by handout #1 titled: P802.3z/D3.1 hot items report. 10:17 HJ done 10:17 DH Del Hanson, the PMD editor, gave a presentation on the progress of the ad hoc Modal Bandwidth Investigation group (MBI). This group was formed at the London interim as a result of comment #754, see below, which affects the fiber link lengths. The MMF link lengths may change to a shorter length. About 25 individuals are involved, having held 4 meeting to investigate the situation, and have set up a work plan. Del presented a slide describing the objectives. The main point is that there is nothing here and now today on this objective/work plan that is complete, and may remain unresolved at time of ballot recirculation. ------------------------------------------------------------------------ Comment #754 Clause 38 Sub-clause 38.10 P38.12, Table 38.8 Comment Type T Ray Lin, Digital Equipment Corp. The Annex 38A physical media dependent link model used to establish link penalties may need to include a differential mode delay (DMD) parameter and measurement specification. Measurements performed at Digital have shown eye pattern closure due to what may be the differences in the differential model delay (DMD) characteristics of multimode fibers not addressed in the link model i. e., as a power penalty. Suggested Remedy Lab measurements will be performed at Digital Equipment Corporation to characterize the DMD parameter relative to 802.3z operation. Preliminary data should be available by the September Interim. Proposed Response The committee directs Del Hanson to bring response/ plan to Santa Clara Meeting. ------------------------------------------------------------------------ 10:25 DH concludes 10:25 HF this is a major issue, and Howard discusses the implications of going forward with this issue unresolved. 10:27 miscellaneous open and spirited discussion from many individuals with pieces as follows: HJ the current document is the best available what are the best/available choices? GOT if it is broken, then we are not able to publish a standard DH knowledge is not yet available RS is it possible to qualify how badly broken the present situation is DH we do not yet have sufficient information to determine 10:30 DH long technical response on model long discussion on what are the work items for the MBI group 10:31 HF Howard presents why this is important to 802.3z, what are the possible range of solutions, and how to reach a decision. One desired action is a tutorial at the Nov. plenary in Montreal. 10:33 DH a long response, we need to quantify just what is happening, good work is in progress. 10:36 HJ this is a complex issue and many people do not understand the issues. HF consider a evening panel discussion at the Nov. plenary DH prefer a more structured and prepared process, tutorial would be better HF the next plenary is too late for a tutorial, by then a decision will need to have been made. DH let’s try for a Wednesday afternoon tutorial (tomorrow) to cover multiple points of views and show scope of problem DH concludes 10:44 HF end of the hot list review break out room & editor assignments Lunch by the pool, hosted by Gigabit Ethernet Alliance ------------------------------------------------------------------------ 802.3z interim meeting minutes, Wednesday Oct. 1, 1997 ------------------------------------------------------------------------ 1:35 HF Howard calls the meeting to order 1:35 GOT presented a pitch for the initial meeting of the 802.3ac working group (Frame Extensions for Virtual Bridged Local Area Networks (VLAN) Tagging on 802.3 Networks) to be held the following day (Thursday, Oct. 1) at Baynetworks. 1:40 HF reviewed where we are at in the agenda, and turned meeting over to Howie HJ hot list review: each editor is to go over sub-working group activities 1:40 BG Clause 6,22,35 editor, Bob Grow reviewed major comments has a response for all comments has a verbal or written sign-off on all comments 1:55 HJ overall activity 4 or 5 comments are expected to remain open for recirculation all editors have blanket approval to incorporate editorial changes SM there are still some not previously circulated comments to cover in Clause 5 comment #541: "Procedure LayerMgmtReceiveCounters is called by the ReceiveDataDecap function in 4.2.9 and increments the appropriate receive counters". change from Reject to Accept comment #542: In the section that deals with lengthError in this procedure, replace all references to LLC with "MAC Client". with whine & whimper from SM, accept the REJECT 2:04 HF Clause 4 editor, Howard Frazer #1 late collisions in 1000 BASE: change transmit status from “transmit error” to “late collision error”. The attempt limit branch then returns to previous status of “excessive collision error”. #2 late collision in IPG: previous action was to only increment the late collision counter. Present action now not only increments counter but also reports count value to layer management. (recording secretary best effort at recreation of Howards statement): Howard emphasized that implementators need to be aware that there is a side effect to this change; the call to layer management is now a different action. With Gigabit, frame and carrier are not the same. Previous, layer management was only called by function TransmitLinkMgnt at the end of a frame/carrier transmission. Now, end of frame is not end of carrier. The call to TransmitLinkMgnt by BitTransmitter for late collision in IPG has no relationship to end of the previous frame. 2:12 HJ Clause 1 editor, Howard Johnson complete 2:12 Clause 30 editor, Sumesh Kaul main activity dealt with remote fault (RF) enumeration: when the PCS informs management about remote fault, then management needs to set the remote fault bits and restart AN to inform the link partner. DL thus management will need write access to the RF bits, and we need to create a comment to cover this requirement. 2:26 Motion #1 extend range of readable/writeable bits for AutoNegotiation Management attributes to include bits D12, D13 Moved: David Law Second: Pat Thaler Technical, > 75% required passes by acclamation 2:28 RT Clause 37 editor, Rich Taborek performed review of new/re-drawn state diagrams Clause 37 is now equal to Clause 28, warts and all major change was to change Next Page from mandatory to optional have proposed responses for all comments Motion #2: accept a (reworded) comment from John Cagle (comment #934) on use of manual configuration while using Auto-Negotiation devices. Adds a new sub-clause 37.1.4.4 as follows: 37.1.4.4 User Configuration with Auto-Negotiation Rather than disabling Auto-Negotiation, the following behavior is suggested in order to improve interoperability with other Auto-Negotiation devices. When a device is configured for one specific mode of operation (e.g. 1000Base-X Full Duplex), it is recommended to continue using Auto-Negotiation but only advertise the specifically selected ability or abilities. This can be done by the Management agent only setting the bits in the advertisement registers that correspond to the selected abilities. 2:30 Moved: Rich Taborek Second: Geoff Thompson Technical, > 75% required Yes: 32 No: 1 A: 5 Passes 2:46 RT complete with PCS, Clause 37 2:46 BG Clause 6, 22, 35 editor, Bob Grow All comments are accepted, rejected, or withdrawn performed a walk thru of GMII electrical 1st pass of D3.2 was used in discussions Clause 22 major editorial to clarify MII/GMII relationship sections and editing instructions renumbered to avoid renumbering (?) 802.3u PICS revised/renumbered Clause 35 major text additions to clarify possible application vs test topology (electrical) clarification of MII/GMII relationship PICS revised/renumbered 2:52 BG done 2:53 RT Clause 36 editor, Rich Taborek Summary: Rich discusses new Annex 36B, which provides calculation examples for 8B/10B transmission code running disparity. At end of discussion, editor is given charter to perform final editing on annex contents (text, table, notes). Summary of discussion follows: BG does this cover reflector discussion PT we need an example with neutral code groups to help cover the general misunderstandings as discussed in reflector 3:00 HF offers a change to example #1 PT provides good words on how error is self healing BG so that there is no multiplication of the error 3:17 RT done 3:17 DH Clause 38 editor, Del Hanson there were a fair number of changes & organization we have keep the same information signal detect is now mandatory there may be operating range changes due to comment #754 HF acknowledges as open and may still remain open as we proceed on to sponsor ballot we need to take a real hard look at this issue in November. not enough information is available today 3:30 DH Tutorial Handout Del Hanson gave a short tutorial on the Modal Bandwidth Investigation which was followed by a spirited Q&A. A summary is: we need more data, we need to quantify the installed base, need to determine how to get from the new DMD method for worst case bandwidth to the old method using ROFL. This is a modeling question for how to screen the fiber. One possible method is to hang a black box or “widget” on the cable. The expectation is that there will be no spec. in Nov. , perhaps choices. Some specific comments during or following the tutorial were: GOT for a plot of DMD vs ROFL, we would expect a straight line DH This has not been done, much work remains HF this is good information, what can we do with it DH in Nov., options are to have quantified: the installed base for new DMD vs old ROFL method the relationship of worstcase bandwidth vs link length DL there is a modeling question for how to screen existing fiber installations PT is there a way to mitigate this problem SS there are possible options, such as hybrid cables 4:07 in Nov., there may be multiple options HF in the general form of launch alignment, special coatings, etc. SS a black box or “widget” to hang on the fiber cable DH do not expect a specification in Nov., perhaps choices 4:12 HJ Slide to propose the following path Installed Base DMD History DMD Worst Case Assumptions ROFL Worst Case Worst Case link distance followed by a spirited discussion of its merits SS each of the above items is open there may be other ways to get there more spirited discussion HJ are we going to be stuck in November? is there a list of possible choices? 4:23 Howie then read into the record the following statement: Requests that Steve Swanson of Corning please honor the customer commitments made in the Coning “fiber you can count on” brochure by providing to our committee the data it needs to do its job. Howie then read into the record the following statement from the Corning brochure: Technical Support: Behind every reel of Corning fiber stands the support of hundreds of technical experts, ready to address any concerns related to optical fiber and its deployment. Corning’s state-of-the-art tracking systems provide answers to specific questions on every reel of fiber produced and purchased. Cornings Center for Fiber Testing, offering more than 60 different tests, is a world leading resource for qualifying more than 60 different test, is a world leading resource for qualifying new product innovations and troubleshooting customer issues. SS provides a very tactful response HJ possible tradeoffs between jitter and length 4:26 DH we do not today know about installed base, for what percentage may be bad (?) target date for information is Nov. plenary. BG what other alternatives are there? DH that is back to the task force to provide. We are not there yet. This problem is also in other fiber. 4:31 DH concludes 4:31 HF Clause 41 editor, Steve Haddock Howard, filling in for Steve Haddock, all open items in those clauses are closed. 4:31 HF we are done with sub-task force break-outs 4:31 HF Agenda item #9, Schedule: Howard discusses the schedule and due to MMF bug, we may slip the schedule. The extreme case of SMF alone with CX is OK. 4:32 HF Agenda item #9, Disposition of comments Motion #3 Accept responses to comments as resolved in subtask force breakout sessions, and reviewed in 802.3z. Produce draft 3.2 and conduct a 20 day recirculation ballot, to close on or about 5 November 1997 , including all unresolved “TR” comments in recirculation package. Allow editors to perform editorial cleanup as required. Recirculation ballot responses will be accepted only via the WWW and e-mail, or hand carried hardcopy. Moved: S. Muller Second: W. Quackenbush Technical > 75% Yes: 36 No: 0 Abstain: 0 Passes 4:42 HF Agenda item #10, Any other business None 4:42 HF Agenda item #11, Plans for next meeting For the Nov. meeting in Montreal, a Monday morning meeting with several small rooms is required 4:48 HF Thanks to GEA for projectors used at this meeting Thanks to David Law and 3COM for London meeting 4:48 HF Agenda item #12, Minutes from London and Maui meetings Motion #4 Accept minutes from Maui and London as written and distributed Moved: Walter Thirion Second: Del Hanson Unanimous via voice approval 4:52 HF we are done, time to adjourn BG Bob Grow DH Del Hanson DL David Law EG Ed Grivna elg@cypress.com GOT Geoff Thompson HF Howard Frazier HJ Howard Johnson PT Pat Thayer RS Rich Seifert RT Rich Taborek SH Steve Haddock SK Sumesh Kaul SM Shimon Muller "'Shimon.Muller@Eng.Sun.COM'" SS Steve Swanson TD Tom Dineen