Monday, November 10, Administrative
Geoff Thompson convened the meeting at 15:00 as IEEE 802.3 chair. Andy Luque, IEEE 802.3 Secretary, recorded the Minutes. Geoff provided a status report on several documents. The IEEE 802.3z, Gigabit, ballot closed August 31. The London interim generated draft 3.2. The recirculation ballot on this draft closed before the plenary. IEEE 802.3aa, Maintenance, is undergoing a recirculation WG ballot. It formally closes on Thursday. A draft of IEEE 802.3ac, VLAN, document was mailed recently. The March 97 Irvine Minutes were distributed a long time ago. However, due to an oversight, the WG did not affirm them at the July plenary.
Geoff reviewed the sign-in procedure and rules for establishing voting rights. Participants introduced themselves. The attendance sign-up list was circulated. The attendance list for the November 10-14 Montreal plenary is included (Attachment A2).
The voting membership of IEEE 802.3 was reviewed (Attachment A2). A list of potential voters was presented (Attachment A2). The procedure has been modified. Participants now need to pick up and sign the new voter form to qualify. David Law will collect the forms. New voters will be identified Tuesday.
Geoff made a call for relevant patents. The patent requirements were reviewed. The IEEE requests release letters from holders of patents that may apply to the standards being developed. These letters should state that the patent holder is willing to comply with the IEEE patent requirements (see example in Attachment A3). The IEEE Standard's Companion explains the policy in more detail. Geoff's policy is to also solicit information on patents that have been filed. It is easier to get release letters, while company representatives are active in the WG.
Several documents were distributed. The Montreal Agenda, March 97 Minutes, July 97 Minutes, and the IEEE 802.3ac VLAN draft were handed out.
The Monday/Tuesday Agenda was reviewed (Attachment A1). There are two possible Agendas for Thursday. We need to select one of the options.
Andy Luque, suggested that we start at 8:00 in the morning. A strawpoll was conducted.
|Start Thursday 8:00||68|
|Start Thursday 8:30||3|
A second strawpoll considered the desire for a lunch break
|No Lunch Break||42|
The Thursday Agenda was adjusted to reflect these votes.
MOTION: Approve the Monday/Tuesday Montreal Agenda
M:Thomas Dineen S: Andy Luque; Approved by acclamation
MOTION: Approve the IEEE 802.3 March 97 Irvine Minutes
M:Howard Frazier S:Andy Luque; Approved by acclamation
The July 97 Minutes will be considered Thursday.
Executive Committee Liaison - Geoff Thompson
Geoff reported on the activities of the IEEE 802 Executive. Several new PARs are under consideration. IEEE 802.5 has several PARs for higher speed projects. IEEE 802.14 has a PAR for a new PHY.
Paul Nikolich indicated that the IEEE 802.14 PAR would allow for a more advanced upstream channel. The new PHY can operate over a noisy channel. It can supports 256 Kb to 20 Mb. It can deal with more hostile environments. He recommends our approval of this PAR.
The IEEE 802 Overview and Architecture PAR is being revised. It eliminates the need for the Functional Requirements. Another WG ballot is planned. IEEE 802.3 liaisons will have an opportunity to review and vote on this document. All comments will be considered.
An IEEE 802.8 document is out for LMSC ballot.
The issue of a reduced fee for students was considered.
The IEEE 802 Executive Committee considered the acquisition of PC projection equipment. However, a ballot on this issue failed. WG input will be solicited later this week. How much do we want to move from a low-tech operation to a high-tech capital intensive group? LCD projectors are expensive. This would be a new level of meeting expense.
IEEE 802.3x and IEEE 802.3y standards will be published in late November. They were not ready for distribution at Montreal.
An election of IEEE 802.3 officers will take place at the March 1998 plenary. Geoff plans to run again. The Vice-Chair and Secretary also need to be affirmed.
The IEEE is still reviewing its book distribution policy. There will be no changes before the March plenary.
Several tutorials are scheduled for Montreal. There will be a tutorial by the Link Aggregation Study Group.
We have until 20:00 Tuesday evening to comment on the PARs.
An announcement on the IEEE 802 Technical plenary will be made tomorrow.
No books will be distributed at Montreal.
The IEEE 802.3 voter requirements were presented. The LMSC Sponsor ballot pool was discussed (Attachment A4). This group reviews our work before it gets final approval from the IEEE. A pool will soon begin to review IEEE 802.3z. November 24 is the deadline for new members that want to participate in the IEEE 802.3z review.
IEEE 802.5 High Speed PARs
The IEEE 802.5 PARs were discussed (Attachment A5). A total of three new PARs are under consideration. They are comparable to our 100BASE-TX, 100BASE-FX, and Gigabit efforts.
Why are there three separate PARs? They did not want to be delay if part of the project felt behind. It gives them an escape clause to publish what is done.
Geoff expressed some concerns. Auto-Negotiation should be required in all future IEEE 802 standards. We need to determine if this is an IEEE 802.3 requirement.
What happens if you attach a High Speed Token Ring device without Auto-Negotiation to an Ethernet Repeater that expects it?
It was noted that Token Ring is a constant transmit media.
Howard Frazier, do the PARs concern DTR, DTX? Geoff, does not think they would object if the work was limited to FD links.
Information was requested on the 5 Criteria. Is there a broad market potential? Geoff, this effort appears to have revitalized attendance at IEEE 802.5. They will probably meet the market potential requirement.
How do we address issues with the PARs? Geoff, they will be discussed Thursday.
It was noted that any comments on the PARs need to made by 20:00 Tuesday evening.
David Law, how do we collate the comments? Make a decision now or form a group to prepare a response.
Howard Frazier, changes need to come from a group. We need a group.
Geoff Thompson, we will form a group to develop an IEEE 802.3 response. Volunteers were solicited. Rich Seifert, David Law, Thomas Dineen, Pat Thaler, Bob Grow and Dan Dove volunteered. Rich will lead the group. They will report Tuesday.
IEEE 802.3w, BLAM - Mart Molle
Mart reported on the status of BLAM (Attachment B). Incorporate the MAC layer changes defined by IEEE 802.3z for half-duplex operation at 1 Gb/s. Developed Gigabit MAC model with and without BLAM. Developed a model of Collision Truncation in half-duplex repeaters. No changes to IEEE 802.3w draft 1.8 (Feb 97). There will be a task force meeting at Montreal. Need to discuss the task force's future role. Recent pricing trends in switching are reducing the market for shared media solutions.
100BASET Maintenance - Colin Mic
Colin reported on the status of 100BASE-T Maintenance (Attachment C). The original ballot closed before the July plenary. Comments were resolved at that time. Changes were made to accommodate existing implementations. They are not impacted. Editorial changes were made to resolve inconsistencies created by the changes made by IEEE 802.3y.
The document is out for a confirmation ballot. No comments have been received as of last night. The ballot closes Thursday. There will be a meeting this week to accept any new change request. The plan is to prepare MB#5 for a LMSC Sponsor ballot.
10/100 Fiber Proposal - John Hill
John announced a BOF session on fiber for Ethernet LANs (Attachment D). BOF session on Wednesday. Will examine market justification for a new proposal. It involves 100BASE-SX, Auto-Negotiation, and SWL (850 nm). At least two companies have expressed an interest. Need 2 hours. Two presentations are scheduled.
RMII Consortium Information - Bill Bunch
Bill provided information on the Reduced MII Consortium (Attachment E). The goal is to reduce the number of pins in the MII from 16 to 6. This is not an IEEE 802.3 project.
Task force room assignments were considered. The meeting was adjourned about 17:00.
Tuesday, November 11, Administrative
The IEEE 802.3 plenary re-convened at 8:30. Geoff Thompson chaired the meeting. Andy Luque recorded the minutes. The attendance list was circulated. Geoff reviewed the voting rules, and voting membership. Eighteen participants expressed interest in becoming voters: Bruce Bandali, Steve Brewer, Steve Dreyer, John Fitzgerald, G.Y. Hanna, Ajit Jadeja, Thomas Jorgensen, Juan Jover, Sumesh Kaul, Arlen Martin, David Nim, Jerry Pate, Tony Rowell, Jim Tatum, Jacob Twersky, Bill Verheggen, Willem Wery, and Mark Yu.
JTC1 Liaison - Geoff Thompson
Geoff provided a report on the activities of JTC1 (Attachment F). The balloting procedures have changed. IEEE 802.3x and 100BASE-T2 are both undergoing simultaneous NP and PDAM ballots. The US Tag has had several teleconferences. The next JTC1 meeting will be in Korea. SC25, SC6 and SC26 will meet in the Far East in the same time frame.
Type Field Assignment - Geoff Thompson
There was a question about the Type field assignments. In response, it was noted that Xerox has transferred this task to the IEEE. Policy rules were developed at the last plenary. These rules will receive final approval sometime this week. There will be a meeting a 14:00 Wednesday. Copies of the rules are available.
All applications for Type fields are to be submitted to IEEE 802. What the group does is still open. Specific numeric assignments will be treated as public information. There are no privacy requirements on either the Type or OUI assignments. The original reason for privacy is no longer valid. This issue can be discussed Wednesday. The fee for each new Type field is $2,500.
1000BASE-T - George Eisler
George reported on the status of 1000BASE-T (Attachment G). The group had an interim meeting in London in September. The Line and Channel codes have been selected. This was the last major issue. The first draft was distributed at Montreal. The group has compiled an initial draft of the Blue Book.
At Montreal, the goal is to complete the choice of a Start-up procedure. Review the draft. Review the Blue Book. The target is a WG ballot after the March plenary. There will be 2 or 3 interim meetings before March.
Geoff Thompson, will there be a tutorial in March. George, yes. However, there may not be enough time. Geoff, we need to bring the electorate up to speed.
Rich Seifert, requested more education.
There was a question about any technical demonstrations of 1000BASE-T. George, after March. Not privy to any demos up to now.
Rich Seifert, concern about voting on a standard that has no physical implementation. George, shares this concern. However, it takes several million dollars to develop a chip. Short of this, it is difficult to prove. Lots of simulations. Rich, that is what we did for 10BASE-T and 100BASE-T. George, you can use discrete components to do 10 Mb/s and 100 Mb/s but you can not do it for 1000BASE-T.
Thomas Dineen, can we use sub-assemblies to demo? George, we can not do it. Partial demonstrations of the technology are not convincing. You can prove it by spending millions to build a chip and then start again once the standard is approved. Demos have an unreal quality. They look OK to the eye but do not prove issues.
Rich Seifert, historically people have spend millions to demo products and then tweaked it when the standard was approved. You do not need to have a standard before you invest. George, does not agree.
Rich Seifert, great difficulty in approving a standard that has not been physically instanciated.
Thomas Dineen, agrees with Rich. These are all individual pieces of work. The issue is systems integration. Can you make all the parts work together?
Pat Thaler, also agrees with Rich and Thomas. We can not do a tutorial and we can not do a demonstration. But, we are suppose to just vote for it. We can not gain perfect knowledge, but we need help understanding the document. George, there are a dozen or so books on the subject. Look at the Blue Book.
Colin Mic, the Blue Book provides a basic discussion of the technology.
Geoff Thompson, we need to put the material in the hands of all IEEE 802.3 voters. This homework is needed before the tutorial in March.
Andy Luque, more of the presentations need to be posted on the web. George, there have been over 200 papers presented. Need someone to set up an archive. More info will be provided on the web.
Gigabit Ethernet - Howard Frazier
Howard reported on the status of the Gigabit Ethernet task force (Attachment H). Several documents were distributed: IEEE 802.3z Montreal Agenda, draft 3.2 ballot comments, an updated draft, and a fiber optics tutorial were distributed. The status of the WG recirculation ballot on draft 3.2 was presented. This ballot closed November 4. There are 5 outstanding ballots that need to be resolved.
All the main technical comments submitted were reviewed. In particular, the unresolved negatives were discussed. The group is still on track for March. A key issue deals with the use of laser on multimode fiber. This concerns the differential mode delay and the link distance of multimode fiber. A total of 1,300 comments have been received during the entire balloting process. The two interim meetings resolved 1,297. A total of 190 were submitted during the recirculation ballot.
Howie Johnson provided more details on the recirculation ballot. There were 16 technical required, 18 technical, and 162 editorial comments. A total of 196 comments.
Scott Mason, ask about the relationship between Auto-Negotiation and Link Start-up. It was noted that this is a carry over item from the WG ballot.
Howard Frazier, we need two MDI connectors. There was a motion in London. A motion on Wednesday will change the requirements. Bob Campbell, rationale was not provided. There was no justification for the two connectors. No standard in IEEE 802.3 has two connectors. It was noted that this issue will be discussed on Wednesday. Howard, 100BASE-FX has 3 connectors. Howie Johnson, look at the relevant discussion from the minutes.
Howie Johnson, the IBM type 1 is not relevant with regard to item #39.
Bob Campbell, plagiarize from Fiber Channel. A connector is to be added. Now a jumper cable. What is the rationale for the note? The note should disappear. Any cable that can meet the I-pattern is usable. We do not need the note.
Howard, we need to know how well lasers work with multimode. We have found some glitches. This is the big hot item of the week.
Item #187, its an open technical issue. How far do we go with multimode fiber?
Item #186, this is issue #754 which was not resolved from the draft 3.1 ballot.
Item #88, Annex 38A, need differential mode delay parameter and measurement specification.
Item #86, jitter budget when using lasers on multimode fiber.
Item #1226, clause 39, TDR requirement on differential balance pair. There is no standard test.
Howard identified the big ticket items. Differential Mode Delay and link distances on MMF is a very big item. A team has investigated. Background and tutorial information has been collected. There will be a special session Wednesday afternoon on PDM. Tutorial and background information in the handout. The first two foils of the tutorial were reviewed.
ATM specified LED and laser sources. LED dominates.
Gigabit improved lasers can modulate faster. LED not fast enough.
Need to control launch.
The current approach is are not satisfactory.
Why did the problem not show up before? IEEE 802.3z is the first to use MMF with gigabit data rates. Fiber Channel is not concerned with the installed base of cable. We have done original research on existing cable plants to try to solve the problem.
Dan Dove's contribution on jitter at 100 meters has been discussed. Chris Di Minico has done some work.
Does new fiber have problem? Howard, it depends on the new fiber. There is no specification to qualify fiber for this problem. It is possible to build fiber to solve the problem.
Launch conditions nailed. Longer distance for MMF are possible.
It could be worst. Just fiber worst not launch conditions.
Del Hansen, we considered worst case overall.
In what way is the launch controlled? the group is working on this.
Field measurements on installed fiber have been conducted. Problems with are in the 1-10% range after testing several labs. At the same time, many links work at longer range than the specification for LX and SX. A more detailed presentation will be made Wednesday.
VLAN Tag Proposal - Howard Frazier
Ian reported on the status of IEEE 802.3 VLAN group (Attachments I). There was an interim meeting October 2. About 50 attended. The group reached consensus on its goals and a PAR was drafted. The first draft was distributed in early November.
The group met Sunday evening in Montreal. The goal is to progress to draft 2. There will be an interim in Seattle. A WG letter ballot is planned after Montreal. The group will meet Wednesday PM if any new issues are raised. Liaisons are being keep with IEEE 802.1 and IEEE 802.5. There will be a page by page review at the Thursday plenary.
IEEE 802.5 High Speed PAR Discussion - Rich Seifert
Rich reported on the review of the 3 new High Speed Token Ring PARs. There was an examination of the 5 Criteria. Questions were raised about the Distinct Identify requirement. FDDI meets the needs at 100 Mb/s. The only difference is the pinout. FDDI allows large frames and priority. IEEE 802.12 already supports native token ring frames. There were no questions with regard to Technical Feasibility or Compatibility. There were questions with regard to Broad Market Potential. No specific numbers were provided.
Several enhancements were suggested with regard to the Purpose and Scope. The use of TX and FX is not specified in the Scope. There is no discussion of the MII. The Economic Feasibility comes into question. Auto-Negotiation is needed for compatibility between 4 Mb/s, 16 Mb/s and 10/100 Mb/s. There are no restrictions on the amount of change that can be made to the MAC.
Howard Frazier, commented on Distinct Identity. The proposed changes would affect the software driver interface.
Geoff Thompson, did you look at what happens when you cross-connect? What type of disruption would occur? Rich, no the group did not look at this issue. Pat Thaler, does the concern relate to Auto-Negotiation? Geoff, more concern about fiber.
Rich Seifert, what is the next step? Howard Frazier, we need to submit something in writing.
Pat Thaler, there is another point that was not covered. Why are the 100 Mb/s fiber and twisted pair efforts divided into two PARs?
Geoff Thompson, we need to add a provision to require all future twisted pair standards to support Auto-Negotiation.
TECHNICAL MOTION: IEEE 802.3 endorses the work done by its Task Force that reviewed the IEEE 802.5 High Speed PARs.
M:Pat Thaler S:Tom Dineen; Y:45, N:0, A:9; Approved
Bob Grow, requested clarification with regard to shared and full duplex goals.
The meeting adjourned.
Thursday, November 13, Administrative
The IEEE 802.3 closing plenary convened at 8:05. Geoff Thompson presided over the meeting. Andy Luque, IEEE 802.3 Secretary, recorded the Minutes. The Thursday Agenda was reviewed. No changes were requested.
The sign-up list was circulated. The voter list was presented. Paul Ahrens, Brad Booth, and Jahan Lotfi were added as new voters.
There was a meeting last night with Bob Love regarding the new IEEE 802.5 PARs.
MOTION: approve the IEEE 802.3 July 1997 Maui Minutes.
M:Andy Luque S: ; Approved by acclamation
IEEE 802.3 Operating Rules - David Law
David reported on the status of the IEEE 802.3 Operating Rules (Attachment J). There were 48 comments on draft 4. Responses were circulated Monday along with draft 4.1. There was meeting Tuesday to deal with 11 comments on draft 4.1. The latest modification to the draft were presented ie. draft 4.2. The text on voting membership was not modified. Balloting groups are established at the close of the preceding WG plenary. Changes to the IEEE 802.3 Rules must be available for preview one week prior to the WG plenary.
TECHNICAL MOTION: adopt the IEEE 802.3 Rules, draft 4.2, as modified, as the Operating Rules for IEEE Project 802 CSMA/CD Working Group 802.3.
M:Thomas Dineen S:Alan Flatman; Y:40, N:0, A:0; Approved
Position on IEEE 802.11 PARs - Geoff Thompson
Geoff indicated that IEEE 802.11 has submitted several PARs for consideration tonight (Attachment K). IEEE 802.11b on Higher speed Wireless LANs, IEEE 802.11c on MAC Bridging and a Maintenance PAR. No one in the WG has provided any input to Geoff on this issue. Expect separate votes on each PAR. Geoff will support these PARs unless directed otherwise.
Rich Seifert, separate Maintenance and MAC Bridge PARs? Yes.
MOTION: vote against the 802.11 High Speed PAR.
M:Rich Seifert S:Andy Luque Y:6, N:20, A:20; Not Approved
The Seifert/Luque motion was debated. Andy Luque, they need to complete work on an Access Point (AP) standard. It is now proprietary.
Alan Flatman, there is a need for more of the same flavor. Moving to higher speed would be a good idea. Do not support Rich's motion.
Rich Seifert, these devices can not interoperate or auto-negotiate.
Geoff Thompson, should we tell them to go around again? Need a better target at the window. This would up the ante.
Pat Thaler, concern about not having notified IEEE 802.11 by Tuesday.
A vote was taken on the Seifert/Luque motion. It was not approved. Another motion followed.
MOTION: support the IEEE 802.11 High Speed PAR.
M:Alan Flatman S:Gary Robinson; Y:17, N:3, A:18; Approved
A vote was taken on the Flatman/Robinson motion. Geoff noted that this was not a directed vote. If they don't meet the 5 criteria, he will not support it. There were no objections to this assertion.
Position on IEEE 802.14 PAR - Geoff Thompson
Geoff indicated that a PAR to improve the speed and robustness of the IEEE 802.14 PHY will be considered tonight (Attachment L). It was noted that the initial IEEE 802.14 standard is still a work in progress.
Rich Seifert, does the new PAR kill the existing one? Tom Dineen, they are trying to respond to a rapidly change market. The market is driven by a consortium. We should support them. The idea is to add value a the top end.
MOTION: support the IEEE 802.14 PAR. M:Tom Dineen S:Rich Seifert; Approved by Affirmation
Position on IEEE 802.5 PARs - Pat Thaler
There was a meeting Wednesday evening to develop an IEEE 802.3 position (Attachment M). At that time, IEEE 802.5 presented an updated PAR and responses to our earlier questions.
In their opinion, Distinct Identify is based on the specific knowledge of implementors and users. They have a different opinion with regard to Compatibility. They are not committed to Auto-Negotiation. If they got a proposal for Auto-Negotiation, they would consider it. In their opinion, Broad Market Potential is demonstrated by the 25 people that are participating from 15 companies.
The Purpose and Scope sections have been improved. They are planning to use the existing MAC.
Are they planning to use FDDI? It refers to 100BASE-TX. The 100BASE-T PMD does not provide significant technical detail, it merely references the FDDI standard. IEEE 802.5 should just reference FDDI directly.
With regard to the multiple PARs for 100 Mb/s, the intent is to start with one draft. If fiber has delays, they can separate the efforts and get approval on what is ready.
Andy Luque, it is not that hard to an extra PAR later if it is needed.
Geoff Thompson, it is very easy to split a PAR. It was done for 1000BASE-T.
Tom Dineen, IEEE 802.5 plans to use frames 10 times larger than those of Ethernet. There may be PHY issues related to the frame buffer and elasticity when larger frames are involved.
Bob Love, will focus on this issue.
Alan Flatman, we need a tutorial on the prime purpose for high speed Token Ring LANs. Why copper vs fiber?
Bob Love, the initial interested is for backbones that are under 100 meters. Many customers want to attach servers.
Geoff Thompson, can you show how much of installed base prefer a fiber backbone.
Bob Love, that is a market issue. Agrees with the use of fiber in backbones.
Rich Seifert, show the 5 Criteria. Concern about the lack of Auto-Negotiation. Are you planning to allow 2 different MACs (802.3, 802.5) on top of a common PHY?
Geoff Thompson, there is no Auto-Negotiation. The pin-out is different.
Rich Seifert, the same cable will carry 4 Mb/s, 16 Mb/s and 100 Mb/s. People support multirate environments.
Bob Love, many users are doing that. It does not require a standard for interoperability.
Pat Thaler, when you leave it up to the implementors, nothing says that there equipment will work with each other.
Rich Seifert, without standard configurations, vendor x and vendor y will not interoperate automatically.
Bob Love, determine the speed.
We need to make a decision to keep on schedule. It was noted that there is no motion on the floor at this time.
Rich Seifert, we already have large frames with FDDI. If this is part of Distinct Identify it should be dropped. We already have FDDI and IEEE 802.12. What is the Distinct Identity for 100 Mb/s Token Ring?
Bob Love, we are missing the leverage of the installed based. IEEE 802.3 has leveraged its existing MAC.
TECHNICAL MOTION: general support for all 3 Token Ring PARs.
M:Hon Wah Chin S:Alan Flatman; Y:23 N:18, A:9; Not Approved
Dan Dove, has not seen general support for these PARs.
Pat Thaler, generally supports how the work is structured. At the same time notes that there are some mistakes. Do we feel strongly enough to delay this work until it is fixed?
Gary Robinson, support the PARs in general. They are the experts in the field. We have the right to make comments. They have the right to accept them. It sets a bad precedent to deny these PARs. If you want to influence the work you need to attend their meetings.
CALL QUESTION: Chin/Flatman motion was called.
Y:55, N:2, A:6; Motion was Called.
The motion was called. A vote was taken on Chin/Flatman motion. The motion was not approved. Another motion was offered.
MOTION: allow the IEEE 802.3 chair to use his discretion with regard to the IEEE 802.5 High Speed PARs.
M:Rich Seifert S:Tom Dineen; Approved by Affirmation
There were no objections to affirming the Seifert/Dineen motion by acclamation. The WG affirmed it by acclamation.
ISO JTC1/SC6 Liaison - Gary Robinson Gary reported on ISO JTC1/SC6 related issues (Attachment N). There has been no additional work since the last plenary. NP are in progress for 1000BASE-T and MB#5. There has been no action on 10 Mb/s.
Gary announced that the ACM Journal is publishing a special issue on IEEE 802. They are looking for articles. Details on how IEEE 802 has made an impact on the industry. Gary is the guess editor. There is a February deadline.
Maintenance 100BASE-T (802.3aa) - Colin Mic
Colin reported on the status of 100BASE-T Maintenance (Attachments O). The WG recirculation ballot closes today. 7 ballots have been returned. No comments. One editorial change. Sponsor ballot will follow if no new comments are received.
TECHNICAL MOTION: forward MB#5 to LMSC sponsor ballot if no new comments are received in WG confirmation ballot. To Revcom once LMSC sponsor ballot successfully completed.
M:Tom Dineen S:Ian Crayford; Y:45, N:0, A:0; Approved
Administrative - Geoff Thompson
Geoff reviewed several administrative issues. The IEEE Executive Committee will be considering hibernation for IEEE 802.2 and IEEE 802.6. Geoff sees no reason to oppose this. IEEE 802.12 may go into hibernation at the next plenary.
The IEEE 802 Executive Committee plans to acquire 5 AV projectors for $3,500 each. The fees will be restructured to reflect this purchase at the July plenary. They will not change for those who pre-register.
VLAN Tag Proposal - Ian Crayford
Ian reported on the status of IEEE 802.3 VLAN group (Attachments P). The task force met Sunday. There were no additional meetings this week. The draft will be updated and a letter ballot will follow. The group will meet at the interim in Seattle. Geoff Thompson, suggested that a provision for a LMSC sponsor ballot be added to the motion to be submitted.
Several liaison have been appointed. Tony Jeffree, Norman Finn, Mick Seaman, Dave Delaney, and Andrew Smith are acting as IEEE 802.1 representatives. Ivan Oakley and John Messenger are acting as IEEE 802.5 representatives. They will part of the WG ballot pool on this standard.
TECHNICAL MOTION: distribute IEEE 802.3ac draft 2 for WG letter ballot by 12/15/97 and also request that IEEE 802.3 authorize IEEE 802.3ac to respond to comments and issue a confirmation letter ballot if necessary. Forward to sponsor ballot if no significant technical changes.
M:Ian Crayford S:Rich Seifert; Y:63, N:0, A:1; Approved
Enhanced MAC - BLAM - Mart Molle
Mart reported on the status of BLAM (Attachment Q). There has been a significant loss of interest in half duplex solutions. There was a task force meeting Tuesday. 6 people attended. The future of the task force was discussed. The task force agreed that the following motion be offered to determine the level of interest in the WG.
TECHNICAL MOTION: IEEE 802.3w/draft 1.8 be submitted to a WG letter ballot.
M:Mart Molle S:John Payne; Y:10, N:20, A:27; Not Approved
Ian Crayford, Mart's work has been excellent. However, people do not want to change the MAC to support half duplex operations.
Alan Flatman, does not see any market relevance. What is the level of vendor commitment?
Mart Molle, a small start-up appears to be interested. IBM is not actively pursuing it. They have ICs but no adapter card. No groundswell of support.
A vote was conducted on the first Molle/Payne motion. It was not approved. A second motion was presented.
TECHNICAL MOTION: IEEE 802.3 return the PAR for IEEE 802.3w to LMSC, with the recommendations that it be forwarded to NesCom for withdrawal.
M:Mart Molle S:John Payne; Y:60, N:0, A:9; Approved
Shimon Muller, how do we preserve the work that has been done? We need some form of achieved to keep this high quality work.
Geoff Thompson, it will be done. Mart has put a tremendous amount of quality work on this project and has always been very gracious.
The motion was amended to affirm IEEE 802.3's appreciation of Mart's significant contribution and dedication to this effort.
A vote was taken on the second Molle/Payne motion as amended. It was approved.
There was break from 10:05 to 10:15.
10/100 Fiber Proposal - John Hill
John reported on the 10/100 fiber proposal (Attachment R). There was a meeting Wednesday. 19 people attended. There was one presentation on market conditions. Identified the 4 highest priority issues. Need extensive investigation before a PAR and 5 Criteria can be drafted. Need to address high priority issues. Another meeting?
MOTION: IEEE 802.3 establishes a Study Group on 10/100 fiber.
M:John Hill S:Tad Szostak; Y:12, N:39, A:19; Not Approved
Andy Luque, is this a technical motion?
Geoff Thompson, no. We only need 50%. We purposely have a low threshold for Study Groups.
The Hill/Szostak motion was debated. How does this differ from existing fiber that goes to the desktop. John, it is shortwave and support Auto-Negotiation. The specific glass has not been determined.
Geoff Thompson, is the proposal to add a new 100 Mb/s PMD to the existing standard? It would support Auto-Negotiation. It would not use plastic fiber but would use existing fiber.
Tom Dineen, this implies another PHY. Why doesn't EIA/TIA take care of this new PHY. John, it is not a characterization of the PHY that we are pursuing.
Tom Dineen, it is a PHY PMD development project. Not fiber.
Howard Frazier, the motion is poorly worded. It is not well defined. It is not a good idea to start a project for which we do not know the target outcome. The level of interest has not been met. Concern that this will only result in significant confusion in the market and dilute resources. 1000 Mb/s Auto-Negotiation is the scariest thing about it. This is a very ill conceived project. We should say no.
Alan Flatman, agrees with Howard. The project is very broad and ill defined. Unable to vote for this.
Pat Thaler, this is suppose to be a pretty low indication that there was interest from a group of people. People are more likely to attend when there has been a motion in favor of a project. It is pretty hard not to pass. There is no count of the number of companies that are interested.
Bill Bunch, calls the question. There were no objections.
Geoff Thompson, the threshold for a Study Group is keep low because we want the freedom to explore, but a vote of support from IEEE 802.3 is needed.
It was noted that other groups also use the 50% criteria.
A vote was taken on the Hill/Szostak motion. It was not approved.
Link Aggregation and Truncking Study Group - Steve Haddock
Steve provided a report on the Link Aggregation and Truncking Study Group (Attachment S). There was a tutorial Tuesday evening. Strong amount of interest. 126 attended the tutorial. A motion to create a Study Group to develop a PAR and 5 Criteria seems appropriate. There will be an interim meeting. Submissions are expected at the March plenary.
TECHNICAL MOTION: IEEE 802.3 form a Study Group for Link Aggregation and Truncking.
M:Steve Haddock S:Tom Dineen; Y:72, N:0, A:2; Approved
The Haddock/Dineen motion was debated. Why do it in IEEE 802.3? Steve, we operate on the premise of a single network. They are aggregated as a single MAC.
Howard Frazier, the intent is to make it suitable for any dedicated IEEE 802.3 link.
Rich Seifert, their is historical precedence for doing it in IEEE 802.3. There is a significant level of interest in IEEE 802.3. Anyone can use it. The people that want to do it are in IEEE 802.3.
Howard Frazier, expect many participants from other WGs. We need a mechanism to insure that others can comment and contribute. This effort would not have a PHY focus. It would preserve the MAC layer as already specified.
Would it be applicable to any media and Switch to Switch? Steve, the time to setup specific goals is in the Study Group.
Howard Frazier, IEEE 802.3 may define different solution for Switch to Switch than to End Systems.
A single MAC makes a lot of sense. There is interest for Switch to Switch solutions. A solution could be implemented in software without changes to the hardware.
Steve, regardless of where the work is done, IEEE 802.1 or IEEE 802.3, it will be the same people.
Howard Frazier, it is not inappropriate to generate multiple PARs if the Study Group recommends it.
Mic Seaman, there is not a lot of value in arguing about turf at the start. When this issue is resolved, participation and voting can be clarified. Steve, all participants get to vote.
Tom Dineen, lets put this issue to rest. In concept, trunks will be implemented in a sublayer to be defined above the MAC and below LLC. No impact to IEEE 802.1.
Rich Seifert, this is not a PAR vote. The Study Group is to elicit input from all WGs.
Shimon Muller, we are doing it here because the interest started in the WG. The first order of business is to set objectives and goals.
Pat Thaler, the concern seems to be Switch to Switch. Presenters are all software vendors.
Floyd Ross, it also involves architecture.
Alan Flatman, we had a tutorial, good background information, and a discussion. Calls the question. There were no objections.
Geoff Thompson, all participants in the room can vote on this motion. A vote was taken on the Haddock/Dineen motion. It was approved. There will be an interim meeting in February in Seattle in conjunction with other IEEE 802.3 task forces. A strawpoll indicated that 64 participants were interested in this effort.
Gigabit Ethernet - Howard Frazier
Howard reported on the status of the Gigabit Ethernet (Attachments T). An update was provided on the resolution of comments from the WG recirculation ballot. The group started meeting Sunday. Comments on the optical PMD have been the key item. As of November 13, the approval rate is 98.59% There are 140 Approves, 2 Disapproves, and 1 Disapprove from an observer.
A total of 196 comments were submitted. One comment requires action from IEEE 802.3. One comment requires follow up with the commenter. Two comments will have to be included in a second recirculation ballot. An updated document, draft 3.3, will be submitted to a recirculation ballot soon after Montreal.
Howard reviewed the main issues reviewed at Montreal. Appropriate TDR test were considered. The Style 2 (HSSDC) connector is now identified as recommended for 1000BASE-CX. The DB-9 connector is retained as an option.
Most significantly, Optical PMD conditioning has been added to mitigate the effects of Differential Mode Delay in MMF with lasers. Consensus was reached on Controlled Launch Conditions. This is a very significant piece of work in terms of the research done. This is a major contribution to the technology. It is useful anywhere MMF is used for high speed communications. The optical PMD distances from draft 3.2 remain unchanged.
Howard solicited questions from the participants. There were none.
Draft 3.3 will be submitted to another WG recirculation ballot. If successful, a LMSC sponsor ballot will follow before the March plenary. There will be a check point at the March plenary before seeking final approval from the IEEE Standards Board. SC6 will receive the draft used in the LMSC ballot. There is a need to get input from 1000BASE-T with regard to jitter on GTX_CLK. Continue work in MBI.
Howard proceeded to review the IEEE 802.3z PAR and 5 Criteria.
Geoff Thompson, indicated an appreciation for the work done by the PMD group. The purpose of the comment is to make sure that if things were not done appropriately, we could not slide around it. This gives the group a chance to review this decision. Howard, believes we have provided appropriate technical information at the Wednesday meeting.
Geoff Thompson, has the cost structure or convenience conditions in this approach affected the Broad Market potential. Howard, the Conditional Launch will not have a material impact on SX. For LX, there is a material cost, a hybrid jumper. Minor component. Also, the added labor cost. These do not affect the cost balance. LX on MMF extends the range but requires an external cable. It preserves the cable plant.
Geoff Thompson, are there any changes to the 5 Criteria? Howard, none.
Howard proceeded to review the PAR and 5 Criteria as requested in Geoff's comment. We met the Hamming Distance requirement. Carrier Extension defined. The 100 meter cooper part was assigned to IEEE 802.3ab. With regard to distance, we support 62.5 um MMF at 260 meters for SX and 440 meters for LX. On SMF, we moved from 2 km to 3 km.
TECHNICAL MOTION: IEEE 802.3 affirms that draft 3.2* as presented adequately fulfills the objectives and 5 Criteria.
M:Tom Dineen S:Alan Flatman; Y:66, N:1, A:3; Approved
*as modified at Montreal 11/97.
It was noted that, if the motion is approved, Geoff's Technical Required comment will be satisfied. The relevant IEEE 802.3z motion was approved by 63/0/0. The Dineen/Flatman motion was approved.
TECHNICAL MOTION: having adopted resolutions for all comments received on draft 3.1 and draft 3.2 of IEEE 802.3z, the Task Force request that IEEE 802.3:
M:Howard Frazier S:Howie Johnson; Y:68, N:0, A:5; Approved
Geoff Thompson, there is no provision if something goes wrong. Howard, suggest another motion.
Andy Luque, offered a friendly amendment. Add to item #4, "with a copy to SC6 for PDAM". It was accepted by the proponents.
What if a new technical issue is found or if comments are made about the changes adopted at Montreal? Will we be able to meet the March goal under those circumstances? Howard, no.
Item #3 deals with editorial issues. What happens if there are technical issues? There is no provision for a second recirculation ballot.
Tom Dineen, how would you view a motion charging IEEE 802.3z, in case of failure, to proceed with "a best effort" within the rules. Howard, we are. The chair has the authority to issue a ballot. It's in the IEEE 802 Operating Rules.
Geoff Thompson, does not want to set a precedent that is dangerous.
Geoff Thompson, the problem is what happens if we find an issue after the IEEE 802.3 plenary. Putting the draft on the IEEE Standards Board's Agenda is not an issue.
How appropriate is it to ask if issues have been created by the changes incorporated into the document? Howard, it is premature to do it in a formal way now.
Andy Luque, called the Frazier/Johnson motion. Jonathan Thatcher objected to calling the question, and requested a vote.
Call Question: call Frazier/Johnson motion. Y:46, N:2 A:12; Called
The question was called. Howard indicated that two people wanted to speak. He requested the indulgence of the chair. Geoff, agreed.
The first person, wanted to know about Jonathan's concern. Jonathan Thatcher, a single vote would take us out of the diagram. If we know of a problem, we could try to addressed it. If anyone has a negative comment please, step forward and raise the issue. We can then try to respond.
A vote was taken on the Frazier/Johnson motion. It was approved.
The LMSC Sponsor voter pool was discussed. Participants were encouraged to join the pool, if they want to participate in the next review level. Please see David Law by the end of today, if you want to be added to the pool.
Del Hansen proposed that a letter be drafted and send to TIA on the Optical PMD issue. It is a follow-up to the Conditioned Laser Launch discussions at Montreal. IEEE 802.3z approved a motion to that effect. IEEE 802.3 was asked to endorsed a similar motion.
TECHNICAL MOTION: IEEE 802.3z request that IEEE 802.3 authorizes the submittal of a letter to be authored by Kolesar/Swanson and approved by the IEEE 802.3 chair (Geoff Thompson) to TIA FO-6.6 (with copies to other appropriate committees) requesting the development of a specification for a higher multimode fiber modal bandwidth with a defined conditioned laser launch.
M:Jonathan Thatcher S:Steve Swanson; Y:33, N:3, A:4; Approved
The Thatcher/Swanson motion was considered. A friendly ammenment to add "MMF" after "laser launch" was accepted. Another friendly ammenment to "send a copy to SC25" was also accepted.
Paul Kolesar, TIA FO-2-2 does not set bandwidth standards. They just deal with measurements. Send the letter to all the affected groups. This is not tied to a change in the bandwidth of the fiber.
Alan Flatman, are we asking for the development of a high bandwidth MM fiber?
Del Hansen, we need to structure it more specifically. Need to move from test method with defined launch condition. FO-2-2 can do it. If we go to increase bandwidth, it goes up a notch. Then we are creating a new category.
Jonathan Thatcher, need another motion to deal with the specification. It does not need to be a motion.
Del Hansen, what if we add a second part?
Howie Johnson, there is no consensus. Suggest we take it off-line and come up with an agreeable proposal. The suggestion was accepted, and the Thatcher/Swanson motion was temporarily tabled.
In the meantime, another motion was considered.
TECHNICAL MOTION: recognizing the need for additional field testing and laboratory work, IEEE 802.3 encourages industry and end user support for these activities, for the benefit of the work of the Gigabit Ethernet Task Force.
M:Howard Frazier S:Tom Dineen; Y:42, N:0, A:3; Approved
The Frazier/Dineen motion was considered. What do you expect from this motion? Howard, to get additional resources.
Paul Kolezar, how are we going to get this support? Howard the letter. Paul, then add that to the motion.
Friendly ammenment to add "to be reflected in letter from task force chair" was accepted. Another friendly amendment to replace "need for" with "benefit of" was also accepted.
A vote was taken on Frazier/Dineen motion. It was approved.
Resumed consideration of the tabled Thatcher/Swanson motion. It was noted that due to the lag time involved in a standard, we are asking for a future item.
Howard Frazier, gigabit is asking for better fiber.
Del Hansen, we keep the specification as written. The object is to extend the distance.
Howard Frazier, this is a future project.
Friendly amendment, add after launch "for future IEEE 802 projects" was accepted.
Friendly amendment, to remove "IEEE 802.3z requests that" was accepted.
Alan Flatman, this offers a migration for SX devices. He has reservations. It will take a lot of time for the work to get done. In several years time. Against the motion.
Geoff Thompson, this is good, but it is a long term fix. In the current scheme, we are using the uncharacterized behaviour of the existing media. Need to characterize existing fiber. It needs to be added to all fiber not just existing fiber. It has been done for copper.
A vote was taken on the Thatcher/Swanson motion. It was approved.
1000BASE-T - George Eisler
George reported on the status of the 1000BASE-T task force (Attachment U). Agreed on a start-up procedure sequence or a blind start-up. Implementors have a choice. There will be an interim in Dallas, January 12-13 and in San Francisco, February 16-17. Plan to request a WG ballot at the March plenary.
A brief tutorial was presented on the 1000BASE-T technology. The goal is to bring the IEEE 802.3 group up to speed. It may not be enough.
Geoff Thompson, is it 250 Mb/s on each pair? Yes, there is a hybrid transformer on each pair. Same as 100BASE-T2. Use 5 level PAM. 100BASE-TX uses 3 level MLT encoding. There are two code points to send each bit. This improves noise immunity. Pulse shaping is used. The viability of these concepts was tested in the 100BASE-T2 silicon.
Andy Luque, requested that more of the 1000BASE-T presentation be placed on the web. George, indicated that this will be done.
There is a Trellis code tutorial in the Blue Book.
The IEEE 802.3 plenary meeting adjourned at 13:36.
Respectfully submitted February 28, 1998.
IEEE 802.3 Secretary
131 McCartney Dr.
Bend, Oregon, 97701
Additional note to WEB version only.
The attachments which form an essential portion of the minutes are only available in paper form. They will be distributed as part of the complete minutes package at the March '98 meeting. They are also available as part of the paper minutes package which can be obtained from the address below.
IEEE Document Distribution Service
Alpha Graphics #35
Attention: P. Thrush
10201 North 35th Ave.
Phoenix AZ 85051 USA
Phone: 602 863-0999
FAX: 602 866-8801