Re: What is so special about POET’s waveguide filtering capability?
posted on
Oct 20, 2019 11:40AM
Just to add a little clarity re Suresh's comment: The same thing is true here: the lasers move.
The lasers do not physically move but the wavelength that the laser emits does move. It shifts based on the operating temperature and characteristics of the laser. Thus it is important that the flat box shape that is represented as the received wavelength have a wide enough range and spacing so that wavelength drift will still be received as the intended channel.
I would also like to add that there are many standards for 400G. But the two that POET is building covers the reach and the advantages that industry requires for data center applications that a Tier 1 customer would want.
IEEE P802.3bs
The table below describes the IEEE Application Acronyms for 400G.
Please note that for formfactor QSFP-DD and OSFP the electrical interface will always operate at 8 x 50G (PAM4).
acronym |
optical lanes |
speed per optical lane |
distance |
wavelength |
comment |
DR |
1 |
100G (PAM4) |
500m |
1310nm |
used for 100G QSFP28 with single lambda working at 112G |
DR4 |
4 |
100G (PAM4) |
500m |
1310nm |
PSM, breakout with DR possible |
DR4+ |
4 |
100G (PAM4) |
2km |
1310nm |
PSM, DR4+ is not a IEEE acronym |
FR4 |
4 |
100G (PAM4) |
2km |
1270, 1290, 1310, 1330nm |
CWDM, no breakout mode |
FR8 |
8 |
50G (PAM4) |
2km |
1270, 1290, 1310, 1330nm |
June 2018: haven't seen on any QSFP-DD and OSFP roadmap |
LR4 |
4 |
100G (PAM4) |
10km |
1295, 1300, 1305, 1310nm |
no breakout mode |
LR8 |
8 |
50G (PAM4) |
10km |
1273, 1277, 1282, 1286, 1295, 1300, 1304, 1309nm |
June 2018: haven't seen on any QSFP-DD and OSFP roadmap |
SR4 2xBIDI |
8 |
50G (PAM4) |
100m |
850/910nm |
400G with MPO12 (instead MPO16). MPO12 is used for 100G SR4 and 40G SR4 |
SR8 |
8 |
50G (PAM4) |
100m |
850nm |
MPO16 required |
Abbreviations:
PSM = Parallel Singlemode
CWDM = Coarse Wavelength Division Multiplex
https://www.flexoptix.net/en/blog/2018/06/the-400g-overview/