Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!zaphod.mps.ohio-state.edu!uwm.edu!mailrus!accuvax.nwu.edu!nucsrl!telecom-request
From: ames!mailrus!uunet!stsusa.com!ellson%sdcsvax@ucsd.edu
Newsgroups: comp.dcom.telecom
Subject: Re: Questions About SONET
Message-ID: <4323@accuvax.nwu.edu>
Date: 23 Feb 90 23:32:15 GMT
Sender: news@accuvax.nwu.edu
Organization: Siemens Transmission Systems, Albuquerque, NM
Lines: 95
Approved: Telecom@eecs.nwu.edu
X-Submissions-To: telecom@eecs.nwu.edu
X-Administrivia-To: telecom-request@eecs.nwu.edu
X-Telecom-Digest: Volume 10, Issue 126, Message 3 of 11


In article <4265@accuvax.nwu.edu>, HGSCHULZ@cs.umass.edu (Henning Schulzrinne)
writes:

> I have a number of questions regarding SONET, the Bellcore-standard
> Synchronous Optical NETwork.
 
> 1. Are there any readily accessible papers (i.e., not just some
> standard) containing details on SONET, beyond the paper in the March
> 1989 issue of the IEEE Communications Magazine? I am especially
> interested in motivations of certain design decisions, not just
> "that's how it is and there is nothing you can do about it".

As a long-time contributor to the development of the Sonet standard I
can probably answer your questions as far as the technical motivations
behind the design decisions in Sonet.

> 2. Why was the row size set to 90 bytes? As it is, ATM packets will
> have to be broken across rows.

ATM was not a consideration when the 9 by 90 frame structure of the
Sonet STS-1 signal was decided, in fact the ATM cell size was not
agreed upon until well after the first release of Sonet.

The 9 row structure was chosen to best accomodate both US and CEPT
digital hierarchies.  3 columns (27 bytes) carries a 1.544Mb signal
and 4 columns (36 bytes) carries a 2.048 Mb signal.

The 90 byte row in the STS-1 signal is sized such that the 9 by 90
frame carries the next major signal in the US hierarchy, the DS3 at
44.736Mb.  The CCITT STM-1 signal, which is equivalent to 3 times the
STS-1, carries the next major signal in the CEPT hierarchy at
139.264Mb (also DS4-NA in the US).

Sonet was designed to reasonably accomodate the all existing digital
hierarchies so that Sonet equipment could be introduced, globally, in
an evolutionary, rather than revolutionary manner.  ATM is expected to
be the first global payload application of Sonet.

The ATM mapping uses the H4 byte to contain an offset indicator to the
next ATM header.  This allows ATM cell alignment to be obtained within
125us of aquiring Sonet frame.  This mechanism would have worked for
any ATM cell size.  I suspect that the availability of this mechanism
was a factor in finally reaching agreement on the cell size; because
the ability of Sonet to carry a particular size was not a factor.

The bytes of the ATM cells arrive sequentially even though they are
broken across rows and frames.  An ATM demapper would simply use a
gapped clock to clock the cell bytes into an ATM queue buffer.
 
> 3. How do ATM and SONET interact? What gets switched where?

Sonet is a circuit switching format, ATM is packet switching.  The ATM
mapping in Sonet permits ATM nodes to be interconnected using
facilities that are shared with the existing digital network.  Sonet
makes extensive provisions for the maintenance of those facilities
thus relieving ATM of the responsibility.
  
> 4. What is the advantage of interleaving ``header'' information
> throughout the frame, rather than concentrating it at the beginning of
> a frame? Why are the payload pointers put a number of rows after the
> beginning of the frame, so that I have to wait until I can determine
> where the payload begins?

I am not quite sure which "header" information you are refering to.  

The ATM cell headers are associated with each cell so that each can be
routed independently.  Remember that adjacent cells may belong to
completely different virtual circuits.

The Section, Line, Path overhead is distributed to minimize the gaps
in the recovered payload clocks that must be smoothed out with
buffering, at least for DS1 and DS3 signals.  (Some would say that the
overhead was not distibuted enough!)
  
> 5. Why was the path overhead made part of the payload rather than the
> header?

The path overhead is not part of the payload, it belongs to the
container that carries the payload end-to-end.  Path overhead is
responsible for ensuring that the payload is succesfully carried
end-to-end across the circuit switched network.  Conversely, Line
overhead only monitors the signal node-to-node, where a node is a
cross-connect or a multiplexer.

In the case of ATM, a Sonet "Path" corresponds to a circuit between 
adjacent ATM nodes.
  
> 6. What is the implementation status of SONET?

Sonet equipment is available now from a number of manufacturers.
            

John Ellson   //   ellson@ontap.stsusa.com   //    602-395-5281
Siemens Transmission Systems, 8620 N 22nd Ave, Phoenix AZ 8502