Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

[EFM] Re: Wavelength allocation

write an EFM standard around CWDM wavelength allocation

I don't think we need to incorporate a wavelength grid into the optical first mile standard.  This is another topic that I feel is out of the scope of EFM work.  The PMDs need to be simplified, not complicated.  Introducing an entire wavelength grid creates several problems:

(1) Box-to-box Compatibility

(2) CWDM does not include 1310 nm, which it seems everybody wants at the ONU
(3) CWDM is not compatible with ITU G.983.3 nor the ITU Grid (used in the Metro)
(4) Admin/Maintenance is a nightmare if each box has a different color laser
(5) We need cheap lasers at the ONU, not expensive wavelength specific or tunable lasers.

A better approach will be, perhaps, to minimize and simplify the PMD wavelength choices, and in at least one case, to leave the 1550 window open (unused).  This way, it is available, but you don't have to decided which is best: CWDM, DWDM, 1550 Analog Video, etc.  

(To give you an idea of the wavelength grid swamp: a while ago I had privy to the WDM plans of several large equipment vendors.  I did a table, side by side, of the wavelengths being used, and ---- they were all different.   In other words, the 8/15/32/etc wavelengths did not match up.  This, wavelength grid inter-compatibility, is a fundamental problem in WDM networks right now. )

In conclusion, I feel a better approach for EFM is to focus on an optical PMD solution that is simple and low cost.   Those that are being studied are:

1550/1310  (P2P, P2MP)
1310/1310  (P2P)
1310+/1310-  (P2P, P2MP)
1490/1310 (P2P, P2MP)

Note that while 1490 nm has not been used before for Ethernet, it is on the CWDM grid, it is specified by ITU-T G.983.3 and it does leave the 1550 window open.

Gerry Pesavento