AFRALTI-TDM Training Workshop on

Network Synchronization, Maputo

Mozambique, November 2011
1. Background (analog telephony, TDM,
PDH)
2. SONET/SDH history and motivation
3. Architecture (path, line, section)
4. Rates and frame structure
1900: 25% of telephony revenues went to
copper mines
standard was 18 gauge, long distance even heavier
two wires per loop to combat cross-talk
needed method to place multiple conversations on a single
trunk

4 kHz
1918: Carrier system (FDM)
5 conversations on single trunk 8 kHz

later extended to 12 (group)

still later supergroups
channels (60), master groups (60)), 12 kHz

f
16 kHz Y(
J)S
SO
NE
20 kHz T
Sli
de
3
Shannon (Bell Labs) proved that
Digital communications
is always better than
Analog communications
and the PSTN became digital
Better means
More efficient use of resources (e.g. more channels on trunks)
Higher voice quality (less noise, less distortion)
Added features
After the invention of the transistor, in 1963 T-carrier system
(TDM)
1 byte per sample 8000 samples per second
T1 = 24 conversations per trunk timeslots Y(
J)S
2 groups per cable! SO
NE

t
T
Sli
de
4
 Analog Crossbar switch Digital Cross-connect (DXC)
1 2 3 4 5 6 7 8
1
1 2 3 4 5
2 t
3
4 processor
5
6 2 1 5 4 3
7 t
Complexity increases rapidly with size
Y(
J)S
SO
NE
T
Sli
de
5
Circuit switching (route is maintained for duration of call)
Route set-up is an expensive operation, just as it was for manual switching

Today, complex least cost routing algorithms are used Y(

J)S
Call duration consists of set-up, voice and tear-down phases SO
NE
T
Sli
de
6
 trunks
circuits

local loop
subscriber line

automatic routing through universal telephone network

Analog voltages used throughout, but extensive Frequency Division Multiplexing

Voice signal arrives at destination after amplification and filtering to 4 KHz

Automatic routing Y(

Universal dial-tone J)S

SO
Voltage and tone signaling NE
Circuit switching (route is maintained for duration of call) T
Sli
de
7
 tandem switch

last mile
PSTN Network
subscriber line
class 5 switch class 5 switch

Analog voltages and copper wire used only in last mile,

but core designed to mimic original situation
Voice signal filtered to 4 KHz at input to digital network
Time Division Multiplexing of digital signals in the network
Extensive use of fiber optic and wireless physical links
T1/E1, PDH and SONET/SDH synchronous protocols
Signaling can be channel/trunk associated or via separate network (SS7)
Automatic routing Y(
Circuit switching (route is maintained for duration of call) J)S
SO
Complex routing optimization algorithms (LP, Karmarkar, etc) NE
T
Sli
de
8
Time Domain Multiplexing relies on all channels
(timeslots)
having precisely the same timing (frequency and phase)
In order to enforce this
the TDM device itself frequently performs the digitization

digital
signals
analog
signals

Y(
J)S
SO
NE
T
Sli
de
9
If the TDM switch does not digitize the analog signals
then there can be a problem
the clocks used to digitize do not have identical
frequencies
we get byte slips! (well, actually, we can get bit slips first
)
exaggerated pictorial
example 1 2 3 4 5 6 7 8 9

1 2 3 4 5 6 7 8 9
component
Numerical example: signals
clock derived from 8000 Hz. quartz crystal 1 2 3 4 5 6 7 8 9

typical crystal accuracy = 50 ppm

So 2 crystals can differ by 100 ppm 1
Y(
2 3 4 5 6 7 9
i.e. 0.8 samples / second
J)S
1 2 3 4 5 6 7 8
TDM SO

So difference is 1 sample after 1 seconds

NE
1 2 3 4 5 6 7 8 T
Sli
de
10
We must ensure that all the clocks have the same frequency

Every telephony network has an accurate clock called

a stratum 1 or Primary Reference Clock
All other clocks are directly or indirectly locked to it (master
slave)

A TDM receiving device can lock onto the source clock

based on the incoming data (FLL, PLL)
For this to work, we must ensure that the data has enough
transitions
(special line coding, scrambling bits, etc.)

1 Y(
J)S
SO
0 NE
T
transitions no transitions Sli
de
11
A clock is said to be isochronous (isos=equal, chronos=time)
if its ticks are equally spaced in time
2 clocks are said to be synchronous (syn=same chronos=time)
if they tick in time, i.e. have precisely the same
frequency
2 clocks are said to be plesiochronous (plesio=near

chronos=time)
if they are nominally if the same frequency
but are not locked
Y(
J)S
SO
NE
T
Sli
de
12
If we want yet higher rates, we can mux together TDM signals
(tributaries)
We could demux the TDM timeslots and directly remux them
but that is too complex
The TDM inputs are already digital, so we must
insist that the mux provide clock to all tributaries
(not always possible, may already be locked to a network)
OR
somehow transport tributary with its own clock
across a higher speed network with a different clock
(without spoiling remote clock recovery)

Y(
J)S
SO
NE
T
Sli
de
13
level

0 64 kbps
* 30 * 24
* 24
1 E1 2.048 Mbps T1 1.544 Mbps J1 1.544 Mbps
* 4 * 4 * 4
2 E2 8.448 Mbps T2 6.312 Mbps J2 6.312 Mbps
* 4 * 7 * 5
3 E3 34.368 Mbps T3 44.736 Mbps J3 32.064 Mbps
* 4 * 6 * 3
4 E4 139.264 Mbps T4 274.176 Mbps J4 97.728 Mbps Y(
J)S
SO
NE
T
CEPT N.A. Japan Sli
de
14
In addition to locking on to bit-rate
we need to recognize the frame structure
We identify frames by adding Frame Alignment Signal

The FAS is part of the frame overhead (which also includes "C-bits", OAM,
etc.)
Each layer in PDH hierarchy adds its own overhead

For example
E1 2 overhead bytes per 32 bytes overhead 6.25 %
E2 4 E1s = 8.192 Mbps out of 8.448Mbps
so there is an additional 0.256 Mbps = 3 %
altogether 4*30*64 kbps = 7.680 Mbps out of 8.448 Mbps
or 9.09% overhead Y(
J)S
SO

What happens next ?

NE
T
Sli
de
15
 digital data rate voice overhead
signal (Mbps) channels percentage

T1 1.544 24 0.52 %
T2 6.312 96 2.66 %
T3 44.736 672 3.86 %
T4 274.176 4032 5.88 %
E1 2.048 30 6.25 %
E2 8.448 120 9.09 %
E3 34.368 480 10.61 %
E4 139.264 1920 11.76 % Y(
J)S
SO

Overhead always increases with data rate ! NE

T
Sli
de
16
analog channels and 64 kbps digital channels
do not have mechanisms to check signal validity and quality
thus
major faults could go undetected for long periods of time
hard to characterize and localize faults when reported
minor defects might be unnoticed indefinitely

Solution is to add mechanisms based on overhead

as PDH networks evolved, more and more overhead was
dedicated to
Operations, Administration and Maintenance (OAM) functions
including:
monitoring for valid signal
defect reporting
alarm indication/inhibition (AIS)
Y(
J)S
SO
NE
T
Sli
de
17
In addition to FAS, PDH overhead includes
justification control (C-bits) and justification opportunity stuffing (R-
bits)
Assume the tributary bitrate is B T
Positive justification
payload is expected at highest bitrate B+T
if the tributary rate is actually at the maximum bitrate
then all payload and R bits are filled
if the tributary rate is lower than the maximum
then sometimes there are not enough incoming bits
so the R-bits are not filled and C-bits indicate this
Negative justification
payload is expected at lowest bitrate B-T
if the tributary rate is actually the minimum bitrate
then payload space suffices
if the tributary rate is higher than the minimum
then sometimes there are not enough positions to accommodate
so R-bits in the overhead are used and the C-bits indicate this Y(

Positive/Negative justification
J)S
SO
payload is expected at nominal bitrate B NE
T
positive or negative justification is applied as required Sli
de
18
 Y(
J)S
SO
NE
T
Sli
de
19
With the disvestiture of the US Bell system a new need arose
MCI and NYNEX couldnt directly interconnect optical trunks
Interexchange Carrier Compatibility Forum requested T1 to solve problem
Needed multivendor/ multioperator fiber-optic communications standard
Three main tasks:
Optical interfaces (wavelengths, power levels, etc)
proposal submitted to T1X1 (Aug 1984)
T1.106 standard on single mode optical interfaces (1988)
Operations (OAM) system
proposal submitted to T1M1
T1.119 standard
Rates, formats, definition of network elements
Bellcore (Yau-Chau Ching and Rodney Boehm) proposal (Feb 1985)
proposed to T1X1
term SONET was coined
T1.105 standard (1988) Y(
J)S
SO
NE
T
Sli
de
20
Rate limitations
Copper interfaces defined

Need to mux/demux hierarchy of levels (hard to pull out a single

timeslot)
Overhead percentage increases with rate

At least three different systems (Europe, NA, Japan)

E 2.048, 8.448, 34.348, 139.264
T 1.544, 3.152, 6.312, 44.736, 91.053, 274.176
J 1.544, 3.152, 6.312, 32.064, 97.728, 397.2
So a completely new mechanism was needed
Y(
J)S
SO
NE
T
Sli
de
21
Synchronous Optical NETwork
Designed for optical transport (high bitrate)

Direct mapping of lower levels into higher ones

Carry all PDH types in one universal hierarchy

ITU version = Synchronous Digital Hierarchy
different terminology but interoperable
Overhead doesnt increase with rate

OAM designed-in from beginning

Y(
J)S
SO
NE
T
Sli
de
22
The original Bellcore proposal:
hierarchy of signals, all multiple of basic rate (50.688)
basic rate about 50 Mbps to carry DS3 payload
bit-oriented mux
mechanisms to carry DS1, DS2, DS3
Many other proposals were merged into 1987 draft document (rate
49.920)
In summer of 1986 CCITT express interest in cooperation
needed a rate of about 150 Mbps to carry E4
wanted byte oriented mux
Initial compromise attempt
byte mux
US wanted 13 rows * 180 columns
CEPT wanted 9 rows * 270 columns
Compromise!
US would use basic rate of 51.84 Mbps, 9 rows * 90 columns
CEPT would use three times that rate - 155.52 Mbps, 9 rows * 270
Y(
J)S
columns SO
NE
T
Sli
de
23
 Y(
J)S
SO
NE
T
Sli
de
24
SONET was designed with definite layering concepts
Physical layer optical fiber (linear or ring)
when exceed fiber reach regenerators
regenerators are not mere amplifiers,
regenerators use their own overhead
fiber between regenerators called section (regenerator section)
Line layer link between SONET muxes (Add/Drop Multiplexers)
input and output at this level are Virtual Tributaries (VCs)
actually 2 layers
lower order VC (for low bitrate payloads)
higher order VC (for high bitrate payloads)
Path layer end-to-end path of client data (tributaries)
client data (payload) may be
PDH
ATM
Y(
J)S

packet data
SO
NE
T
Sli
de
25
 ADM regenerator ADM
Path Line Section Line Path
Termination Termination Termination Termination Termination

path
line line line
section section section section

SONET (SDH) has at 3 layers:

path end-to-end data connection, muxes tributary signals path section
there are STS paths + Virtual Tributary (VT) paths
line protected multiplexed SONET payload multiplex section
section physical link between adjacent elements regenerator section

Each layer has its own overhead to support needed functionality

Y(
SDH terminology J)S
SO
NE
T
Sli
de
26
A SONET signal is called a Synchronous Transport Signal
The basic STS is STS-1, all others are multiples of it - STS-N
The (optical) physical layer signal corresponding to an STS-N is an
OC-N

SONET Optical rate

STS-1 OC-1 51.84M
STS-3 OC-3 155.52M *3
STS-12 OC-12 622.080M *4
Y(
STS-48 OC-48 2488.32M *4 J)S
SO
NE
STS-192 OC-192 9953.28M *4 T
Sli
de
27
 Y(
J)S
SO
NE
T
Sli
de
28
framing

Synchronous Transfer Signals are bit-signals (OC are optical)

Like all TDM signals, there are framing bits at the beginning of the
frame Y(

However, it is convenient to draw SONET/SDH signals as rectangles SO J)S

NE
T
Sli
de
29
 90 columns
framing
9 rows

Each STS-1 frame is 90 columns * 9 rows = 810 bytes

There are 8000 STS-1 frames per second
so each byte represents 64 kbps (each column is 576 kbps) Y(
Thus the basic STS-1 rate is 51.840 Mbps J)S
SO
NE
T
Sli
de
30
 270 columns

9 rows

Synchronous Transport Modules are the bit-signals for SDH

Each STM-1 frame is 270 columns * 9 rows = 2430 bytes
There are 8000 STM-1 frames per second
Y(
Thus the basic STM-1 rate is 155.520 Mbps J)S
SO

3 times the STS-1 rate!

NE
T
Sli
de
31
 SONET SDH columns rate
STS-1 90 51.84M
STS-3 STM-1 270 155.52M
STS-12 STM-4 1080 622.080M
STS-48 STM-16 4320 2488.32M
STS-192 STM-64 17280 9953.28M

STS-N has 90N columns STM-M corresponds to STS-N with N =

3M
SDH rates increase by factors of 4 each time
STS/STM signals can carry PDH tributaries, for example: Y(

STS-1 can carry 1 T3 or 28 T1s or 1 E3 or 21 E1s

J)S
SO
NE
STM-1 can carry 3 E3s or 63 E1s or 3 T3s or 84 T1s T
Sli
de
32
 SONET SDH T1 T3 E1 E3 E4
STS-1 28 1 21 1
STS-3 STM-1 84 3 63 3 1
STS-12 STM-4 336 12 252 12 4
STS-48 STM-16 1344 48 1008 48 16
STS-192 STM-64 5376 192 4032 192 64

E3 and T3 are carried as Higher Order Paths (HOPs)

E1 and T1 are carried as Lower Order Paths (LOPs)
(the numbers are for direct mapping) Y(
J)S
SO
NE
T
Sli
de
33
 90 columns
3 rows
9 rows

Synchronous Payload Envelope

6 rows

Transport Section overhead is 3 rows * 3 columns = 9 bytes = 576 kbps

Overhead framing, performance monitoring, management
TOH
Line overhead is 6 rows * 3 columns = 18 bytes = 1152 kbps
protection switching, line maintenance, mux/concat, SPE
pointer
Y(
SPE is 9 rows * 87 columns = 783 bytes = 50.112 Mbps J)S
SO
Similarly, STM-1 has 9 (different) columns of section+line overhead
NE
T
! Sli
de
34
 270 columns

RSOH

MSOH

Section
Overhead
SOH
STM-1 has 9 (different) columns of transport overhead !
RS overhead is 3 rows * 9 columns
Pointer overhead is 1 row * 9 columns Y(

MS overhead is 5 rows * 9 columns J)S

SO
NE
SPE is 9 rows * 261 columns T
Sli
de
35
 9*N
columns

9 rows

270*N columns

3 STS-1s can form an STS-3

4 STM-1s (STS-3s) can form an STM-4 (STS-12)
4 STM-4s (STS-12s) can form an STM-16 (STS-48)
etc. for STM-N (STS-3N)
Y(
J)S
SO

The procedure is byte-interleaving

NE
T
Sli
de
36
...

Y(
J)S
SO
NE
T
Sli
de
37
SONET/SDH receivers recover clock based on incoming signal
Insufficient number of 0-1 transitions causes degradation of clock
performance

In order to guarantee sufficient transitions, SONET/SDH employ a scrambler

All data except first row of section overhead is scrambled
Scrambler is 7 bit self-synchronizing X7 + X6 + 1
Scrambler is initialized with ones

A short scrambler is sufficient for voice data

but NOT for data which may contain long stretches of zeros

When sending data an additional payload scrambler is used

modern standards use 43 bit X43 + 1
run continuously on ATM payload bytes (suspended for 5 bytes of cell tax)
run continuously on HDLC payloads
Y(
J)S
Xn Yn = Xn + Y n-43
SO
NE
T
Sli
Z-43 de
38
 A1 A2 J0 The STS-1 overhead consists of
section 3 rows of section overhead
overhead
B1 E1 F1 frame sync (A1, A2)
D1 D2 D3
section trace (J0)
error control (B1)
H1 H2 H3 section orderwire (E1)
Embedded Operations Channel (Di)
B2 K1 K2 6 rows of line overhead
pointer and pointer action (Hi)
line
D4 D5 D6 error control (B2)
overhead D7 D8 D9 Automatic Protection Switching signaling
(Ki)
D10 D11 D12
Data Channel (Di)
Synchronization Status Message (S1)
S1 M0 E2 Far End Block Error (M0)
line orderwire (E2)
Y(
J)S
SO
NE
T
Sli
de
39
 m
A1 A1 A1 A2 A2 A2 J0 res res
media
RSOH dependent
B1 m m E1 m F1 res res
(defined for
SONET radio)
D1 m m D2 m D3

AU pointers
B2 B2 B2 K1 K2 res
reserved for
D4 D5 D6 national use

MSOH D7 D8 D9
D10 D11 D12
S1 M1 E2
Y(
J)S
SO
NE
SOH T
Sli
de
40
A1, A2 - framing bytes
A1 = 11110110
A2 = 00101000

SONET/SDH framing always uses equal numbers of A1 and A2

bytes

J0 - regenerator section trace (in early SONET - a counter called C1)

enables receiver to be sure that the section connection is
still OK
enables identifying individual STS/STMs after muxing
J0 goes through a 16 byte sequence 1 C1 C2 C3 C4 C5 C6 C7
MSBs are J0 framing (100000)
0 S S S S S S S
Cs are CRC-7 of previous frame Y(

S are 15 7-bit characters J)S

SO

section access point identifier 0 S S S S S S S

NE
T
Sli
de
41
B1 Byte Interleaved Parity-8 byte
even parity of bits of bytes of previous frame after scrambling
only 1 BIT-8 for multiplexed STS/STM

E1 section orderwire
64 kbps voice link for technicians
from regenerator to regenerator

F1 64 kbps link for user purposes

D1 + D2 + D3 192 kbps messaging channel

used by section termination as Embedded Operations Channel
(SONET) Y(
J)S
or Data Communications Channel (SDH) SO
NE
T
Sli
de
42
In SONET, pointers are considered part of line overhead
For STS-1, H1+H2 is the pointer, H3 is the pointer action
H1+H2 indicates the offset (in bytes) from H3 to the SPE
(i.e. if 0 then J1 POH byte is immediately after H3 in the row)
4 MSBs are New Data Flag, 10 LSBs are actual offset value (0 782)
When offset=522 the STS-1 SPE is in a single STS-1 frame
In all other cases the SPE straddles two frames
When offset is a multiple of 87, the SPE is rectangular

To compensate for clock differences

we have pointer justification
When negative justification
H3 carries the extra data
When positive justification
Y(
J)S

byte after H3 is stuffing byte SO

NE
T
Sli
de
43
If tributary rate is above nominal, negative justification is needed
When less than 8 more bits than expected in buffer
NDF is 0110
offset unchanged
When 8 extra bits accumulate
NDF is set to 1001
extra byte placed into H3 H1 H2 extra
offset is decremented by 1 (byte)

If tributary rate is below nominal, positive justification is needed

When less than 8 fewer than expected bits in buffer
NDF is 0110
offset unchanged
When 8 missing bits
NDF is set to 1001

byte after H3 is stuffing H1 H2 H3 stuff
Y(
J)S
offset is incremented by 1 (byte) SO
NE
T
Sli
de
44
B2 BIP-8 of line overhead + previous envelope (w/o
scrambling)
N B2s for muxed STM-N

K1 and K2 are used for Automatic Protection Switching (see

later)

D4 D12 are a 576 Kbps Data Communications Channel

between multiplexers
usually manufacturer specific OAM functions Y(
J)S
SO
NE
T
Sli
de
45
S1 Synchronization Status Message
indicates stratum level (unknown, stratum 1, , do not
use)

M0 Far End Block Error

indicates number of BIP violations detected

E2 line orderwire
64 kbps voice link for technicians
from line mux to line mux Y(
J)S
SO
NE
T
Sli
de
46