[AusNOG] Acceptable dBM loss over SM cross connect in a datacentre

[Ross Wheeler]

> Hi Ross,
>
> Power loss, reported as -19dBm.

That doesn't make any sense to me. Perhaps "the industry" has adopted some 
perculiar term compared to RF, but "loss" doesn't have a reference.

Expressing loss as a negative seems counter-intuitive too, so a "loss of 
minus 19dB" is a double-negative, suggesting a gain of 19dB (seems 
unlikely in a passive lump of glass!)

Power will be given in dB from some known reference, eg:
dB(m) (dB referenced to 1 milliwatt)
dB(V) (dB referenced to 1 volt)
dB(W) (dB referenced to 1 watt)
etc...

+30dBm = 1 watt.       (10^3 * 1mW = 1W)
-20dBW = 10 milliwatts (10^-2 * 1W = 0.01W)

To give a "loss" of -19dBm sounds like someone doesn't really understand 
what they're telling you, and they either mean "19dB loss", or "we're 
measuring -19dBm of power". While many people use the terms interchangably 
(wrongly) they then rely on you to interpret it correctly from the 
context.

(Much like people getting kW and kWh mixed up and using the wrong ones)


> area where they've been plugging SC cables into SCA sockets), I don't have
> photos, but the power loss is now measuring -12dBm.  Which my switches now
> report as normal.

Your switches are (probably) reporting the receive POWER at -12dBm (which 
would sound about right). I doubt the switch any way to know what power 
the other end is sending and therefore no way to know what the attenuation is.
(a 100km cable might have 10dB attenuation, transmit power of -3dBm would 
see a receive power of -13dBm. The same transmitter on the end of a short 
jumper cable would report close to -4dBm)