Re: 664b/66b frame alignment in SONET payload
Erik,
In addition to the valid sync locations, I am concerned about doing another self synchronizing data stream inside of a synchronous
transport. This is what PPP does and it has a data loss model of up to 8%, depending on the size and loading of the network. For
all practical purposes, all of that data loss is at the port interface between the SONET and the Network Layer switch matrix. These
are reported as "dropped frames". I have attempted to determine the cause of the data loss but can not find anything determinable.
SONET links have a BER of 10e-14 or better, so data loss that great is not because of the synchronous transport. There has been a
suggestion that HDLC is the major cause of errored frames, but SDLC and HDLC have been around for a long time before PPP and were
not considered the cause of inherent errors. The only thing that I can not quantify is the effect of the "gear box" that is used
to map the self synchronizing data stream into a synchronous transport. Without anybody having operational experience with 64B/66B,
I am even more concerned.
Thank you,
Roy Bynum
----- Original Message -----
From: "Erik Trounce" <erikt@xxxxxxxxxxxxxxxxxx>
To: <stds-802-3-hssg@xxxxxxxx>
Sent: Wednesday, June 14, 2000 2:31 PM
Subject: 664b/66b frame alignment in SONET payload
>
> Hi,
>
> I have a question regarding alignment of the PCS 66-bit frames
> within the SONET payload (in WAN mode obviously). My understanding
> is that the PCS 66-bit frames can and will cross SONET frames, and
> obviously they will not begin on octet boundaries. However, will
> it be specified that the PCS frames will sit on even-bit boundaries,
> reducing the valid sync locations from 66 to 33?
>
> Erik
>
>