Karen,
While I have heard discussion of breakout to 16 lanes of 25GbE from a module like CDFP, I propose that it would be a very small niche. Breakout is not an objective
in 802.3bs.
The CDFP was designed to support 13 CDFP in a 19” switch to enable 5.2Tb/s of throughput (13 X 400G) while a QSFP28 can support 1.8Tb/s in one row (18 X 100G),
so CDFP can provide about three times the bandwidth density in a single row. The high density of the electrical channels of the CDFP makes belly-to-belly or stacked configuration very challenging, so it is difficult to scale CDFP to more than one row of modules
to deliver 10.4 Tbps (2 X 13 X 400G) in 1U. The QSFP28 on the other hand is commonly used in a belly-to-belly or stacked configuration to yield 3.6Tb/s (2 X 18 X 100G).
Kind regards,
Scott
From: Karen Liu [mailto:karen@xxxxxxxxxxxxx]
Sent: Monday, December 22, 2014 8:47 AM
To: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-3-400G] 400G Electrical Channels
There’s an assumption below that optics dominates the module size. It’s unclear to me that’s going to remain true. I think this comes from additional implicit
assumptions from the way things have been done to date, maybe about linear arrays. This seems akin to assuming optics will always be NRZ. As we go forward, the optics just have to get a little bit cleverer. Electronics isn’t built all in linear layout.
With some clever layout, we think should be possible to get dense 16-lane modules even for single-mode. A key metric from an application standpoint is face-plate
density. CDFP isn’t allowing the 2x, 4x, etc desired growth in power density per 19” rack. Is there a need for a narrower (physically) electrical interface that would support SR16 breakout to 25G?
Best,
Karen
From: Ali Ghiasi [mailto:aghiasi@xxxxxxxxx]
Sent: Friday, December 19, 2014 12:14 PM
To: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-3-400G] 400G Electrical Channels
Scott
On a separate email to John/Joel/Vasu I raised the concern defining 50G signaling to address CDAUI-8 applications could potentially tie our hand when defining the high volume cost sensitive 50 GbE based on 1x50G (LAUI). It will be more
constructive to define 50G signaling in the project that also defines 50 GbE, a clear set of application requirements potentially could drive the consensus something we lack now.
As far as the real 400GbE application 2km/10 km
the optics will dominate the module size, a narrower interface will not immediately help. On the other hand the only PMD the group had consensus to date to adopt has been 400Gbase-SR16 based on CADUI-16, a CADUI-8 interface will force 400Gbase-SR16
to use an inverse mux/gearbox of some sort. Considering break out applications, with 25 GbE standardization under way the right interface will be CDFP with CDAUI-16 and not a CDAUI-8 interface.
Based on the current dynamic of the group, I am not sure we will reach consensus in the bs project on the 50G signaling with both NRZ and PAM4 looking viable to address CADUI-8 applications. During Q3-14 OIF probably took the right approach
by accepting both NRZ and PAM4 as bases for the 50G XSR/VSR/MR interfaces.
Thanks,
Ali Ghiasi
Ghiasi Quantum LLC
On Dec 19, 2014, at 8:04 AM, Scott Kipp <skipp@xxxxxxxxxxx> wrote:
400G Task Force,
In the electrical ad hoc yesterday, I objected to using the materials loss tool because I thought it could be misused. After Beth explained how people should use the tool, I could not retract my objection before the conversation went out of control. I agree to use the tool to predict losses, but would like for people to use the tool appropriately.
I would like to build consensus to what kinds of electrical channels this project should support. The question that arises is:
Should we define the same electrical channels for 1X50G, 4X50G and 8X50G when they are serving different markets and volumes?
I propose that 8X50G (CDAUI-8) electrical channels should be different than channels that will eventually be defined for 1X50GbE (LAUI) because:
1. Crosstalk - Crosstalk will impact the channel by more than 3dB. I base this off some excellent work that Avago presented to T11 earlier this month. In 14-409v0, Avago showed how going from 1X28G (32GFC) to 4X28G (128GFC) caused a 2.5dB decrease in the channel mainly due to crosstalk in the QSFP28 connector. I would expect more crosstalk than 2.5dB when the speed and number of channels are doubled.
2. Small market - The 400GbE market based on 802.3bs standards will be small compared to single lane 50GbE channels that won’t be standardized for a few years. 802.3bs needs to recognize that this project is targeting a very small market for high-end router interconnects. The complete router market was less than 2M ports/year according to leading analyst firms and not growing much. 16X25G and 8X50G will never be deployed in millions of ports/year because it will be made obsolete by a future 4X100G 400GbE project. 50GbE will go after a larger market and be designed into high-volume, server interconnects by the end of the decade while 4X100G technology will be in the next decade.
3. High cost – Because of the low volume of 8X50G and the high quality required by the router market, the cost will be high for 400GbE. Because of the premium nature of the router market, electrical channels will be made of high quality, high cost materials that can be absorbed in high end router products.
4. Better Materials – The group should model channels based on the better materials supported by the high-end router market in 2018. As others said, the materials being targeted in 802.3bs should be different than the materials targeted for high volume servers in 802.3bj and a future 50GbE project.
Based on these 4 factors that are related, we need to agree on a distance that needs to be supported and consider crosstalk and the loss/inch of better materials. Based on these difference, I bet that CDAUI-8 electrical channels will be 5dB different than the eventual LAUI-1 electrical channels.
We should make sure that CDAUI-8 is optimized for our application, not for LAUI-1 that might be optimized for server and switches.
I’d like to hear if you disagree with these claims.
Kind regards,
Scott Kipp