| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
Hi Mike
I can certainly add the 100G cases to this slide deck – it just gets unwieldy in terms of the diagrams (# lanes) and so I tried to keep it simple – as Chris stated I hoped it would be an easy extrapolation from the 50G to 100G cases by multiplying the number of lanes involved by 2 for each part of the link.
In either the 50G or 100G case, the stated backwards compatibility goal is to take advantage of evolving efficiencies at the physical later (i.e. going from 2 x 25G to 1 x 50G or 5 x 25G to 2 x 50G) – and yet be able to use the higher speed physical layer with legacy 25G based hardware boxes. As I have tried to make sure I always stated, this may lead to different SNR budgets depending on where you terminate the PCS – however we haven’t had a detailed analysis of that yet.
Thanks
Rob
From: Chris Cole [mailto:chris.cole@xxxxxxxxxxx]
Hi Mike
One quick way to convert the 50G cases to 100G is change single lanes to a double lanes, for example LAUI-2 to CAUI-4.
From: Mike Dudek [mailto:mike.dudek@xxxxxxxxxx]
On slide 5 of the reference it says “ Note: 50GE cases taken for simplicity, applies to 100GE as well”
It isn’t clear to me how this applies to 100G as the legacy system would be 100G. I think an equivalent slide showing what 100G would look like (maybe including discussion of module form factors) would be helpful.
Mike Dudek QLogic Corporation Director Signal Integrity 26650 Aliso Viejo Parkway Aliso Viejo CA 92656 949 389 6269 - office.
From: Chris Cole [mailto:chris.cole@xxxxxxxxxxx]
Hi Brad,
We tried to capture the details of the various applications in Rob’s presentation:
http://www.ieee802.org/3/50G/public/adhoc/archive/stone_021716_50GE_NGOATH_adhoc-v2.pdf
We will have an updated presentation in Macau.
More broadly, you raise a very important question which is the relevance of IEEE 802.3 broad market potential, economic feasibility, and distinct identity criteria to the realities of the market place today. The trend has been for exponential proliferation of optical PMDs and form factors. At 10G the industry consensus was that lowest cost is achieved by reaching agreement on a couple of solutions (SR and LR in SFP+) which everyone focused on developing and deploying, hence creating a broad market. For 100G and higher, the new industry approach to lowest cost is developing unique technology optimized to individual applications. MS nicely articulated this perspective in 802.3 by commenting that it will deploy whatever best solves its problems, independent of whether it’s standardized in the IEEE or not. Other Web 2.0 companies have similar approach. Since different end users have different requirements, the industry has responded by developing multitude of different solutions. Which raises the question whether there will even be such a thing as a broad market in the optics industry, and if not then it’s problematic to use it as a criteria for new standards.
Chris
From: B Booth [mailto:bbooth@xxxxxxxx]
If we're going to use titles like "Supported Applications" can we at least highlight what those applications are. Where in the network is that link intended to be used? What is the intended application space or market? From: Wangxinyuan (Xinyuan) Another option? Just borrow from Chris table, with “S3”.
发件人: Chris Cole [mailto:chris.cole@xxxxxxxxxxx]
The 50/100/200G SG will be making a choice in Macau between two objective setting scenarios for 50, 100, 200Gb/s xAUI chip-to-module interfaces.
[The entire original message is not included.]
|