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

Re: A Minor NIT

At 10:34 PM 6/21/00, Thomas Dineen wrote:

>         Wesley has a point for an STS-192c payload. Please
>note small c for SONET concatenation.
>         Sorry if I confused anyone with my original note, of this morning,
>which was referring to and STS-1 payload.
>         As STS-192 payload, concatenated or not, features 3 * 192 = 576
>columns of overhead, and 192 * 87 = 16,704 columns of payload
>for a total of 17,280 columns, and of course 9 rows.
>         For concatenated payloads, such an STS-192c, the pointers
>must also be multiplied, by 192 in this case, to allow them a radius
>which covers the span of the entire payload.
>         So yes pointer adjustments do occur in increments of
>192 bytes for an STS-192c frame! But how?
>         For an STS-192c a positive pointer adjustment will add 192
>stuff bytes to this SPE, and thus delete 192 payload bytes from this SPE.

No. The 192 bytes inserted as part of a positive stuff event are not considered as part of the SPE. During a positive stuff the 192 bytes after the H3 bytes are stuffed (with all 0's I think) and are not considered as part of the SPE. During this period of time the incoming SPE bytes are backup up in a small elastic store. On the 193rd  byte after the H3 bytes SPE data begins to be read out of the elastic store again and inserted into the SONET frame. 

At the receiving end SPE data is extracted from the SONET frame and also written into an elastic store. The receiver knows that during a positive stuff event that the 192 bytes following H3 are stuff bytes and therefore does not write them into the elastic store. 

>A negative pointer adjustment will delete 192 stuff bytes, and thus add
>192 payload bytes to this SPE. These stuff bytes are stored in each of
>the 576 H3 bytes in the Line Overhead. Please note that in either case
>all of the payload bytes get there, eventually.
>Thomas Dineen
>Rich & Tom,
>I had thought I understood pretty much everything, but now on the 
>issue of pointer adjustment, there may be a problem.  As I understand
>this, if there is a positive adjustment, there is need to to
>byte stuff (or is it 192 * bytes?).  But since the SPE is composed of 
>66-bit code words, how can we stuff the SPE by the 192 bytes. Pls
>correct me if my understanding is wrong.
>-Wesley Lee
>Very well said. You echo my understanding also. This implies that SONET/SDH
>pointer adjustment, while operating on byte boundaries, is 100% compatible
>an SPE payload that is on 66-bit boundaries. Does anyone out there disagree
>this statement?
>Best Regards,