| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
Thank you all for a great call yesterday! Below are my own personal notes, sharing for those that couldn’t make the call. Please send me any pictures of applications that you have permission to include! Thank you again and see you in
Berlin with a new draft to discuss. ------------------------------------------------------------------------------------------------- NEA – 100G Electrical Signaling Ad Hoc call – 6/29/2017 IEEE Patent Policy shown IEEE Participation Policy shown Agenda approved without opposition 2 main goals for the meeting: Discussion on the compatibility of optics CFI preliminary draft slide deck (goergen_100cfi_nea_01a_0617) – Does each slide have value? What’s missing? Soliciting feedback. Discussion on the compatibility of optics Strong desire to be compatible and support the already existing DR specifications *Title change to include multi-lane AUI for DR4 ALL existing optics should be considered, not pulled in – Pete brought up to keep future looking FECs/Optics in mind “Should we bring in DR in?” is a loaded question.
à Be careful in determining the scope for the study group - for CFI request AND wording for motion.
ß The intention of the question was to get on the same page as far as compatibility. Stating AUI will imply that we’re considering the optics/FEC. Would not want to go into Backplane discussion without the option to discuss FEC. Want to have a 100G-based AUI… Final Take Away: Language will be critical to distinguish what is being included and what is not. What was the intent of the backplane here? Should “front plane” copper cable also included? 100G backplane PMD should consider shorter traditional, orthogonal, and cabled backplane, but not to exclude “front panel” cables Suggestion: PMDs for backplanes and Cu Cable – include pictures of a variety of backplanes and cables AI – connector vendors – looking for cabled pictured (front-end, chip-to-chip, backplane, etc) Do we want to include onboard optics?? Or leave to OIF (XSR) and COBO It’s just another AUI, isn’t it? – table this discussion to the study group Including AUI in the CFI, allows us to discuss in the Study Group AI – If you have a picture of COBO that you’re safe to use – submit it – we’ll put it in the deck to show wide variety of applications **Remember that the CFi’s goal is to show that there is consensus on the fact there is work that needs to be studied! Will the new rules for length (<= 1yr from the first meeting) of the study group affect the ability to have these full discussions?? David law says “NO!” – Will help us through this if need be. John points out that the CFI deck is really just proof of and tool for building consensus… however, the actual CFI is the request and the vote. So perhaps a verbal CFI would be sufficient if we need to extend a Study group. CFI draft slide deck (goergen_100cfi_nea_01a_0617) Typical outline: Market Drivers, Technical Feasibility, and Why Now?
ß this pattern seems to drive consensus Market Need Content Show multiple things driving this Market Need… show ALL of the applications…
Recommendation: start with the next logical step, here’s where is fits into the family, “go faster to go denser to grow” Look at 400G CFI deck for good slide about the “balancing act” needed between tradeoffs of density and cost. We (as an industry) already agree on the broad market potential! Nothing has changed in the market potential, but we are just changing the technology and need to rebalance the solution for this speed. Suggestion for bulk of the Market Need section: include slides 16, 21, 13, 12 (or 14) Making the case for the need of 200/400G isn’t needed… just lean on the density aspect of serial Use slides 13-14 to tie in the number of interfaces in the system and the number of interfaces Technical Feasibility Content 2 approaches: Layer stack/architecture or implementation Tyco, Molex,, Credo, and Maycom have already presented good content for this area Need to craft the story: This is what the SerDes can do, these are the channels that can be built, system vendors can live with something like this – just keep it practical and plausible… show there is a possible story out there.
The system vendors aspect is covered in broad market potential. Here’s a plausible story, but look at the options in the toolbox that could be used to better the sorry! Be careful on wording… Chip-to-fabric raised a discussion of what fabric means… Consider a slide emphasizing diversity of implementation – Backplane = Internal longer reach channel or broader concept (NOT “printed circuit backplane” as
in .3bj)
Other Topics for study group – a detailed list isn’t needed in the CFI, but in the initial study group discussions best to put together these lists and topics Will need to keep in mind of error propagation – limitation of FEC discussion for the study group. “… for existing PHYs” Attendees (taken from the WebEx list): Adam Healey Adrian Butter Alan McCurdy Ali Ghiasi Dale Murray Dave Lewis David Chalupsky David Law David Malicoat David Piehler Dino Pozzebon Frank Chang Gary Nicholl Geoff T Howard Heck James Fife Jane Lim Joel Goergen John D’Ambrosia Kent Lusted Marcus Sun Megha Shanbhag Mike Dudek Nathan Tracy Oded Wertheim David Ofelt Phil Sun Piers Pirooz Tooyserkani Rajmohan Hegde Rich Mellitz Rick Rabinovich Robert Lingle, Jr. Tom Palkert Yang Zhiwei Beth Kochuparambil |