Kitz Forum
Chat => Tech Chat => Topic started by: Weaver on January 09, 2022, 11:31:27 PM
-
I found this good Cisco article about GPON protocols (https://www.cisco.com/c/en/us/support/docs/switches/catalyst-pon-series/216230-understand-gpon-technology.html#anc0). The article mentions 30 + 15 = 45 bytes of overheads for GPON and GEM, but that figure seems to depend on a parameter (which they term N) that I don’t understand at all. I also need to find out about the higher levels of the protocol stack at that point, and I would also like to look into the overheads and protocol stack for FTTP further upstream than GPON, across the path to the ISP. Can anyone help me out? (Apologies if I have asked about this before - probably have :) ).
-
(Apologies if I have asked about this before - probably have :) ).
I don't recall you previously asking about this topic. :) However I will not be able help. :(
-
Not familiar with the overheads at PHY layer but happily they don't impact the customer experience unless the PON is congested.
Customers are rate limited at Ethernet layer either via GPON's shapers or the bandwidth of the Ethernet port.
The G.984.4 Implementer’s Guide and the standard itself should be an interesting read.
-
Hey Weaver :)
That 'N' is a variable field - the upstream MAP tells ONTs/ONUs when to transmit and for how long. If there isn't much demand on the upstream on the PON there will be fewer 8 byte MAPs sent at each MAP period. N just indicates however many 8 byte blocks are needed to provide all the TDMA timeslots for all ONTs. A PON with 16 devices on it that are barely using upstream will not be very populated while one where the split is saturated by 64 ONTs transmitting small frames will need more blocks.
Another overhead mentioned there but not elaborated on is FEC. This is optional on GPON, required on XGSPON.