I attend IEEE 802.3 Ethernet standards meetings and blog about them from time to time. ( For past blog posts, see the list at the bottom of this post ). The most recent 802.3 meeting was held in Pittsburgh and has just finished. Pittsburgh is very interesting with fine buildings and nice parks. It sits at the confluence of two wide rivers and was a major transportation and steel town in its day. I've included some photographs which I took while there. Before telling you about what happened at the meeting I'll briefly explain the process by which Ethernet standards are made. There is a base Ethernet standard currently known as 802.3-2012. Every few years, a revision is made to the base standard to include all the new standards that have been published since the last revision. The next revision, IEEE 802.3-2015, will be published soon and will incorporate the recently published 100 Gb/s 802.3bj and 802.3bm standards. The last two letters of the standard designate the project that amended the base standard. “bm” for example refers to the P802.3bm project that added 100GBASE-SR4 100G 4-lane optics and the CAUI-4 electrical interface. “bs” refers to the project to add PHYs for 400G speeds. Project naming runs from a-z, then from ba-bz, and then from ca-cz. P802.3bz has just got underway to standardize 2.5GBASE-T and 5GBASE-T PHYs. The next project will be 802.3ca. Before a new Ethernet standard is published, it has to go through a number of gates. The sequence is as follows: There is a “Call for Interest” at an 802.3 meeting. If the call for interest is successful a study group is formed to make a formal request to the IEEE standards association to start a new 802.3 project. The study group will meet during 802.3 meetings to determine its objectives and create its “Project Authorization Request.” If this is successful, a task force will be formed to work on the project. The first job of the task force is to adopt technical proposals to meet the project’s objectives. To be adopted, a technical proposal must be voted in by 75% of the people attending the task force meeting. Sometimes consensus is reached fast, for example in the 802.3bz project, and sometimes slowly, for example in the 802.3bs project. Once a task force has adopted its technical proposals, it can create the first draft of the standard, usually numbered 1.0. This will be reviewed and revised until it is technically complete. At this stage it is offered to the wider 802.3 “working group” as draft 2.0 for approval (ballot) and comment. If the 802.3 working group approves the draft, it will be further commented on and revised until it is offered to the IEEE standards association for approval as draft 3.0. At his stage the standard goes through further cycles of comment and revision before being finally published. For the process to work, people have to turn up to meetings, submit technical proposals, reach sufficient consensus to agree things with a 75% majority vote, volunteer to act as editors of the draft standards and officers of the task forces, and vote and make comments against the draft standards. It is quite remarkable the process works at all, and it is much to the credit of the volunteer 802.3 leadership that it does. At Pittsburgh the 802.3bz task force held their first meeting and much to their surprise adopted technical proposals to meet all their objectives. Their editor, George Zimmerman, who is also responsible for drafting the 802.3bq 40GBASE-T standard, will now have to work overtime to create the first draft standard for 2.5GBASE-T and 5GBASE-T in time for the next meeting in July. These new PHYs are needed to connect to wireless access points for the new 802.11 standards that require greater than 1 Gb/s data rates. 10GBASE-T PHYs are not suitable as they cannot use the legacy cable infrastructure. Also in Pittsburgh after many meeting cycles of deadlock, the 802.3bs task force finally agreed on the modulation that should be used to meet their objective of 400G over SMF optical fibre: Use 8 wavelengths (lambdas) of PAM-4 to for the 10km SMF objective (one fibre in each direction, 50G per lambda) Use 4 fibres in each direction of 100G PAM-4 to support 500m SMF objective (it will be a challenge to develop 100G serial technology) The 2 km objective will probably be dropped as it is covered by the 10 km objective. At the next meeting in July, the 802.3bs task force is expected to agree on whether to go with a 1 by 400G or a 4 by 100G FEC solution. When this is done they will have adopted all their technical proposals and will finally be able to create the first draft of 802.3bs. Another thing of note from the Pittsburgh meeting is that the 802.3bw standard for 100BASE-T1 automotive 100 Mb/s PHY is now successfully through standards association balloting and can be expected to be published soon. The 1000BASE-T1 gigabit automotive PHY 802.3bp standard is making good progress and will go to ballot by the 802.3 working group after the next meeting in July. So there is much interest in developing specialist Ethernet PHYs for automotive applications. Also of relevance to automotive applications is the 802.3bv project which is developing a gigabit PHY for use over plastic optical fibre. The 802.3bv draft standard will also go to working group ballot after the next meeting. Personally I am a member of the 802.3by editorial team which is writing a standard for 25 Gb/s PHYs for MMF, twinax cable and backplane. It was a busy meeting for us resolving comments against the drat 1.0 standard. This project is also making good progress with the next draft expected to be good enough to go to working group ballot after the next meeting. If you want more information about IEEE activities you can find it here: http://www.ieee802.org/3/ Cadence has a comprehensive design and verification IP portfolio for Ethernet and we strive to keep it current to the latest standards. Arthur Marris Related stories: -- Highlights from Recent IEEE 802.3 Ethernet Standards Meeting -- Get a Glimpse at New Ethernet Standards in the Works -- 400G Task Force, 100G Backplane Project and Other Highlights from IEEE 802.3 Ethernet Standards Meeting -- Latest Developments in Ethernet Standards -- Great Progress with Ethernet Standards Development![]()
![]()
↧