TECHNICAL TUTORIAL

Subject: SDH

Date: October, 2003

Prepared by: John Rumsey

___________________________________________________________________________

SDH
Synchronous Digital Hierarchy

1. Introduction
2. The Plesiochronous Digital Hierarchy (PDH)
3. The Synchronous Digital Hierarchy (SDH)
4. Conclusion

1. Introduction
This document provides a technical discussion of SDH.

It explains the differences between the Plesiochronous Digital Hierarchy and the Synchronous Digital Hierarchy.

Before we start it is important that you understand the difference between Synchronous, Plesiochronous,
Isochronous and Asynchronous.

These are all different ways of synchronising a data stream between the transmitter of the data stream and the
receiver.

They all refer to how a data stream is clocked.

Synchronous (Synchronised)
All of the clocks are synchronised to a master reference clock. They may be out of phase with each other but
they will run at exactly the same frequency.

Plesiochronous (Almost Synchronised)

All of the clocks run at the same frequency to a defined precision. These clocks are not synchronised to each
other so the data streams will run at slightly different rates.

Isochronous (Synchronised)
An Isochronous data stream has the timing information embedded in it (eg. a G.704 stream). These data streams
can be carried over Synchronous or Plesiochronous networks.

Asynchronous (Not Synchronised)

The clocks are not synchronised. The transmitter and the receiver have independent clocks that have no
relationship with each other.

1
2. The Plesiochronous Digital Hierarchy (PDH)
In a PDH network you have different levels of Multiplexers.

Figure 1 shows three levels of multiplexing:- Plesiochronous Digital Hierarchy

2Mbit/s to 8Mbit/s 2Mbps
2Mbps
8Mbit/s to 34Mbit/s 8Mbps

34Mbit/s to 140Mbit/s
34Mbps

So to carry a 2Mbit/s data stream across the 140Mbps

140Mbit/s trunk requires it to be multiplexed up
through the higher order multiplexers into the
140Mbit/s trunk and then to be multiplexed down
through the lower order multiplexers.
Figure 1
Because Plesiochronous is not quite Synchronous
each of the multiplexers need a little bit of
overhead on their high speed trunks to cater for the slight differences in data rates of the streams on the low
speed ports.

Some of the data from low speed ports (that are running too fast) can be carried in the trunk overhead, and this
can happen at all multiplexing levels. This is known as Justification or Bit Stuffing.

2.1 PDH Multiplexing Hierarchy

Figure 2 shows that there are two totally different hierarchies, one for the US and Japan and another for the rest
of the world.

The other thing to notice is that the different PDH Multiplexing Levels
multiplexing levels are not multiples of each other.

For example CEPT2 supports 120 Calls but it United States & Japan Europe & Australia

requires more than 4 times the bandwidth of Multiplexing

Name # calls
Rate
(Mbps) Name # calls
Rate
(Mbps)
Level
CEPT1 to achieve this. 1 DS1 24 1.544 CEPT1 30 2.048

2 DS2 96 6.312 CEPT2 120 8.448

This is because PDH is not exactly synchronous 3 DS3 672 44.736 CEPT3 480 34.368
and each multiplexing level requires extra 4 DS4 4032 274.176 CEPT4 1920 139.264
bandwidth to perform Bit Stuffing.

So the Plesiochronous Hierarchy requires Bit Figure 2

Stuffing, at all levels, to cater for the differences
in clocks.

This makes it particularly difficult to locate a particular 2Mbit/s stream in the 140Mbit/s trunk unless you fully
de-multiplex the 140Mbit/s stream all the way down to 2Mbit/s.

2.2 Drop & Insert a 2Mbit/s stream

To drop & insert a 2Mbit/s stream from a The PDH Multiplexing Mountain
140Mbit/s trunk you need to break the 140Mbit/s
trunk and insert a couple of 34Mbit/s to
140Mbit/s multiplexers. You can then isolate the
140Mbps 140Mbps
appropriate 34Mbit/s stream and multiplex the
other 34Mbit/s streams back into the 140Mbit/s
34Mbps
trunk.
8Mbps
Then you de-multiplex the 34Mbit/s stream, isolate
the appropriate 8Mbit/s Stream and multiplex the 2Mbps 2Mbps

other 8Mbit/s streams through the higher layer

Figure 3
multiplexer, into the 140Mbit/s trunk.

2
Then you de-multiplex the 8Mbit/s stream, isolate the 2Mbit/s Stream that you have been looking for and
multiplex the other 2Mbit/s streams up through the higher layer multiplexers into the 140Mbit/s trunk.

Figure 3 shows the multiplexer mountain required to drop & insert the 2Mbit/s stream.

2.3 The limitations of PDH:-

PDH is not very flexible
As previously explained, it is not easy to identify individual channels in a higher order bit stream.
You must multiplex the high rate channel down through all multiplexing levels to find a particular lower speed
channel. This requires an expensive and complex multiplexer mountain.

Lack of Performance
It is not easy to provide good performance if you cant monitor the performance in the first place.
For PDH there is no international standard for performance monitoring and no agreed management channels.
There are some spare overhead bits that are being used for management but they have limited bandwidth and are
hard to locate in a 140 Meg stream without de-multiplexing.

Lack of standards
Not only does PDH have two totally different multiplexing hierarchies but it is quite weak on standards. For
example there are no standards for data rates above 140Mbit/s and no standards for the line side of a Line
Transmission Terminal.

3. The Synchronous Digital Hierarchy (SDH)

SDH, like PDH is based on a hierarchy of continuously repeating, fixed length frames designed to carry
isochronous traffic channels.

SDH was specifically designed in such a way that it would preserve a smooth interworking with existing PDH
networks.

The developers of SDH also addressed the weaknesses of PDH. They recognised that it was necessary to adopt
not only a Synchronous frame structure but one that also preserves the byte boundaries in the various traffic bit
streams.
SDH Multiplexing Hierarchy
Because SDH is synchronous it allows single stage
Data Rate SONET SDH
multiplexing and de-multiplexing. This eliminates (Mbps) (USA) (Europe)
hardware complexity. You dont need multiplexer 51.84 STS-1, OC-1 (not defined)
STS =
mountains. 155.52 STS-3, OC-3 STM-1 Synchronous
Transport Signal
466.56 STS-9, OC-9 STM-3

3.1 SDH multiplexing levels 622.08 STS-12, OC-12 STM-4 OC =

Optical carrier
Figure 4 shows the SDH multiplexing levels. 933.12 STS-18, OC-18 STM-6
1244.16 STS-24, OC-24 STM-8 STM =
The US and Japan use SONET while most of the Synchronous
1866.24 STS-36, OC-36 STM-12 Transport Module
rest of the world use SDH. 2488.32 STS-48, OC-48 STM-16
9953.28 STS-192, OC-192 STM-64
Apart from using some different terminology, there Figure 4
is very little difference between SONET and SDH.

You can see that the data rates are the same except SDH doesnt specify a 51 Meg rate.

STM-1 forms the basis of the SDH frame structure. For example an STM-4 is a frame consisting of 4 x STM-1s.

In Sonet, the STS levels refer to the speed of the bit stream. When these bits are converted to a train of optical
pulses in a fibre, they are called an Optical Carrier (OC).

You may also see OC-3c referred to. This is simply the same bit rate as OC-3, but interpreted as one channel
instead of 3 multiplexed OC-1s. The c stands for Concatenated.

3
3.2 SDH Network Elements
Figure 5 shows the elements that make up an SDH network.

Path Terminating Element

These are the end point devices where the lower SDH Network Elements
speed channels enter and leave the SDH Network. E1

These are known as Path Level devices. E1

PTE
STM-n STM-n
REG
STM-n
E1

Path Terminating Regenerator

Digital Cross Connect Element

These devices can x-connect at the STM level STM-n

down to individual E1 streams. So an E1 stream on E1 E1 E1

STM-n STM-n
one STM trunk could be x-connected to another STM-n STM-n
DCS
ADM
STM trunk. Digital Cross
Connect
STM-n
Add/ Drop Mux

Regenerator Figure 5
Is a device that regenerates the signal.

Add/Drop Multiplexer
The Add/Drop mux has the ability to breakout and insert low speed channels into an STM stream.

SDH Network Configurations SDH Network Configurations

Point to Point Point to Multi-Point

E1 E1 E1
E1 E1 E1 E1
E1 E1 E1 E1
PTE STM-n STM-n PTE STM-n STM-n STM-n STM-n
E1
REG
E1 PTE REG REG PTE
E1 ADM E1
Path Terminating Regenerator Path Terminating
Element Element
Path Terminating Path Terminating
Element Add/ Drop Mux Element

Figure 6
Figure 7

3.3 SDH Network Configurations

The simplest network configuration is a Point to Point network as shown in figure 6. This involves two terminal
multiplexers linked by fibre with of without a regenerator in the link.

SDH Network Configurations SDH Network Configurations

E1 E1 E1 E1
Mesh Architecture STM-n E1 Ring Architecture
PTE
E1
REG
ADM STM-n
E1 Digital Cross Path Terminating Element E1
Connect
E1 E1 Add/ Drop Mux
PTE STM-n STM-n STM-n STM-n PTE E1 E1
E1
REG
DCS REG
E1
E1 ADM ADM E1
Path Terminating STM-n Path Terminating
E1 E1 E1
E1 E1
Element E1 REG
Element
Add/ Drop Mux Add/ Drop Mux
E1
PTE STM-n ADM
E1

Path Terminating Add/ Drop Mux

Element Figure 8 Figure 9

If we include an Add/Drop mux we can now have a point to multi-point configuration as shown in figure 7.

A meshed network architecture (as shown in figure 8) uses a Digital Cross Connect to concentrate traffic at a
central site and allow easy re-provisioning of the circuits.

4
The most popular network configuration is the Ring Architecture shown in figure 9. Here we have four
Add/Drop muxes interconnected by 2 fibre rings. The main advantage of this architecture is its survivability.
If a fibre is cut or an Add/Drop mux dies the multiplexers have the intelligence to heal the network.

3.4 The SDH Frame

The basis of SDH is the STM-1 Frame as shown in figure 10.

The STM-1 frame runs at 155.52Mbit/s, and is

125uS long. This means that you get 8,000 STM-1 The Basis of SDH Frame Structure
frames per second.
STM-1 Frame STM-1 Frame STM-1 Frame S
8,000 frames a second is a very common rate in
telecommunications networks for example G.704 125uS
operates at 8,000 frames a second.
155.52 Mbit/s
Payload
This means that each Byte in the frame is equal to a 8,000 Frames per Second Section Overhead
64kbit/s channel.
1 Byte = One 64kb/s Channel

The Frame is made up of a Section Overhead Figure 10

field and a Payload field.

STM-1 Frames are usually represented as 9 Rows by 270 Columns for a total of 2430 Bytes as shown in figure
11. The bytes are transmitted from Left to Right, Top to Bottom.

The first 9 Columns are the section overhead and

the other 261 columns are used to carry the STM-1 Frame Structure
payload. 270 Columns (2430 bytes)

The Section Overhead has three parts:-

* Regenerator Section Overhead Section Virtual Container 9

Overhead Capacity Rows
* Pointers
* Multiplex Section Overhead

In SDH the actual user data is carried in Virtual

9 Columns
Containers. The Virtual Containers have a Path Figure 11
Overhead field and they come in a number of
different sizes.

We will look in detail at Virtual Containers later in the tutorial.

But first we will have a look at the SDH Overhead.

3.5 SDH Overhead

Figure 12 shows how the SDH overhead and transport functions are divided into the following layers:-

Regenerator Section
The Regenerator Section Overhead contains information required for the elements located at both ends of a
section. This might be between two Regenerators or Line Termination Equipment.

5
 Multiplex Section
The multiplex section overhead contains
information required between the multiplex section SDH Overhead
termination equipment at each end of the Multiplex Path
Section. Basically that means between consecutive Multiplex Section Multiplex Section
network elements excluding Regenerators.
Regenerator Section Regenerator Section

Path
The Path overhead is created and terminated by the E1 E1
PTE STM-n STM-n STM-n STM-n PTE
Path Terminating Equipment at either end of the REG ADM REG

link. It is transported in the Virtual Container with Path Termination

Section Multiplex Section Section Path Termination
Termination
the user data. Termination Termination

Figure 12

In SONET terms these are known as Path,

Line and Section. And sometimes these terms are incorrectly used for SDH.

Figure 13 shows the structure of the overhead bytes.

The first 3 rows are Regenerator Section Section Overhead

Overhead:-

A1 and A2 are framing bytes and indicate the start Regenerator A1 A1 A1 A2 A2 A2 J0

of the STM-1 frame. Section
Overhead
B1 E1 F1
D1 D2 D3
Pointers H1 H1 H1 H2 H2 H2 H3 H3 H3
J0 contains a trace message that is continually B2 B2 B2 K1 K2
transmitted between Regenerator Sections so that Multiplex D4 D5 D6
Section D9
they know they are still connected. Overhead
D7 D8
D10 D11 D12
S1 E2
B1 provides parity checking. Calculated over all
bytes of the previous STM-1 frame. Figure 13

E1 can be used for voice communications over the regenerator Section

F1 is set aside for the users purposes.

D1, D2 and D3 form a 192kbit/s message channel for Operations, Administration and Maintenance. Eg. Alarms,
control, monitoring.

The last five rows of the Section Overhead are used for the Multiplex Section:-

B2 24 bit Parity Check

K1 and K2 Automatic Protection Switching. This is used to provide automatic switching to a backup facility in
the event of a failure.

D4 to D12 form a 576kbit/s message channel for Operations, Administration and Maintenance. Eg. Remote
provisioning, monitoring, alarms etc.

S1 is the synchronous status message byte used to carry synchronisation messages ie: I am primary reference
clock or Do not use me as clock reference.

M1 Remote error indication

E2 A 64kbit/s voice channel between Multiplex Sections.

6
Path Overhead
The Path overhead is part of the Virtual Containers. Figure 14 shows the overhead associated with high order
and low order Virtual Containers.

This overhead is transported between Path

Path Overhead
Path
Terminating Devices:- Overhead

J1 Virtual Container
High Order Virtual Containers B3
VC-3 48.960 Mbit/s 85 Columns
J1 This byte repetitively transmits a user C2
VC-4 150.336 Mbit/s 261 Columns
G1
programmable data string so the receiving path F2
device can see that it is still connected to the H4
V5 Virtual Container
intended transmitting path device. J2 VC-2 6.912 Mbit/s 12 Columns
F3
N2 VC-12 2.304 Mbit/s 4 Columns
K3 VC-11 1.728 Mbit/s 3 Columns
K4
B3 Parity bits N1

Figure 14
C2 Specifies the type of mapping in the Virtual
Container. For example it tells you if the High Order Virtual Container contains Lower Order Virtual Containers
or ATM CELLs etc.

G1 Conveys performance of the path

F2 Communications between path elements

H4 Multiframe indicator

Low Order Virtual Containers

There is only a single byte of overhead in a low order Virtual Container. This byte is used for the following
purposes over a 4 byte multiframe.

V5 Error Checking, Path Status and Signal label (async, byte sync, bit sync)

J2 Access path identifier (so the receiver knows he is still connected to the intended transmitter)

N2 Connection Monitoring

K4 Automatic Protection Switching. This is used to provide automatic switching to a backup facility in the event
of a failure.

3.6 Virtual Containers

Figure 15 shows the 5 Virtual Container sizes and Virtual Containers
the services they are intended to carry.
Virtual
Container Bandwidth For carriage of
It is a very rigid structure and is not very
VC-4 150.336 Mbit/s CEPT4 (139.264 Mbit/s)
bandwidth efficient.
VC-3 48.960 Mbit/s DS3 (44.736 Mbit/s)
CEPT3 (34.368 Mbit/s)
For example to carry an E3 (CEPT3) service
requires more than 30% overhead.
VC-2 6.912 Mbit/s DS2 (6.312 Mbit/s)
Even the best case an E1 requires more than 10%
VC-12 2.304 Mbit/s CEPT1 (2.048 Mbit/s)
overhead.
VC-11 1.728 Mbit/s DS1 (1.544 Mbit/s)
And there is no provision to carry N x 64kbit/s, E2
Figure 15
(8Mbit/s) or 10 meg or 100 Meg.

The Virtual Containers provide a permanent nailed up path across the network and there is no Dynamic
Bandwidth Allocation.

7
 Multiplexing Structure
3.7 SDH Multiplexing Structure
xN x1
Figure 16 shows how to combine multiple services STM-N AUG AU-4 VC-4 C-4 139264k

onto the STM trunk. x3

x3
TUG-3
x1
TU-3 VC-3

44736k
If we look at the example of a E1 (2Mbit/s) service. AU-3 VC-3 C-3
34368k
x7
x7 x1
C-2 6312k
1. The E1 frame is placed into a C-12 Container Pointer Processing
TUG-2 TU-2
x3
VC-2

2. A Path overhead is added and it becomes a VC- Aligning TU-12 VC-12 C-12 2048k

2 Virtual Container. Multiplexing

Mapping
x4
VC-11 C-11 1544k
3. Multiple VC-12s are assigned Pointers and TU-11

become a TUG-2 Tributary Unit Group. The Figure 16

pointers indicate the location of the first byte of

each of the Virtual Containers. STM-1 with VC-4
4. Seven of these TUGs can be Mapped into a VC- 270 Columns (2430 bytes)
3 Virtual Container
1
5. Multiple VC-3 Virtual Containers will be 2
Regenerator
Section
assigned Pointers and placed into an AUG 3
Overhead
Virtual

Path Overhead
Administration User Group 4 Pointers Virtual Container
Container 9
6. And the AUG will be placed in the STM Frame. 5 Capacity Rows
6
7
Multiplex
Section VC-4
Overhead 150.336 Mbit/s
The Pointers are used to locate individual 2 meg 8 (149.760 Mbit/s)
9
streams in the STM Frame.
9 Columns
Figure 17
3.8 Virtual Container Examples
Figure 17 shows how a VC-4 Virtual Container fits
into the STM-1 Frame.
STM-1 with 3 x VC-3s
The VC-4 fits perfectly in the STM-1 Frame and a 270 Columns (2430 bytes)
Pointer indicates the location of the first byte of the 1 Regenerator
VC-4. 2 Section
Overhead
3
48.960 Mbit/s

48.960 Mbit/s

48.960 Mbit/s
Path Overhead

Path Overhead

Path Overhead
If the VC-4 is not properly synchronised with the 4 Pointers Virtual Container
VC-3

VC-3

VC-3
9
5 Capacity Rows
STM-1 frame it can slip position in the Frame. 6 Multiplex
7 Section
Overhead
There is actually a Byte of bandwidth in the Pointer 8
9
Section that can be used by the VC-4 if it is
9 Columns
running faster than the STM-1 Frame. Figure 18

Figure 18 shows how 3 x VC-3s fit into the STM-1 Frame.

Although it is not shown in this diagram the

VC-3 Virtual Containers do not take up all the STM-1 with 3 x VC-3s
bandwidth of the STM-1. They are allocated a 270 Columns (2430 bytes)
certain fixed space and they can move around in 1
Pa
48
48.960 Mbit/s 48.

Regenerator
Pat

it. There is extra bandwidth if the VC-3s are 2 Section

Overhead
running from faster clocks than the STM-1 3
48.960 Mbit/s

48.960 Mbit/s
Path Overhead

Path Overhead

Path Overhead

Frame. 4 Pointers Virtual Container

VC-3

VC-3

th OverheadVC-3

9
5 Capacity Rows
6 Multiplex
.960 Mbit/s

The pointer indicates the first byte of the Virtual

960 Mbit/s

Section
h Overhead

7
Overhead
VC-3

Containers. 8
VC-3

Figure 19 shows how the VC-3 Virtual 9 Columns

Figure 19
Containers may start in different places within
their assigned area in the STM-1 Frame.

When a device places a Virtual Circuit into the STM-1 Frame it also sets the Pointer Value so the receiving
device can locate the beginning of the Virtual Circuit.

8
Figure 20 shows how Lower Order Virtual STM-1 with 3 x VC-3s with VC-12s
Containers are Carried in the STM-1 Frame:- 270 Columns (2430 bytes)

1
To hold Lower Order Virtual Containers we need 2
Regenerator
Section
to use a TUG (Tributary Unit Group). Each TUG 3
Overhead

Path Overhead

Path Overhead
has a fixed place in the Virtual Circuit and it has a

48.960 Mbit/s
48.960 Mbit/s
4 Pointers
number of Pointers in fixed positions in the TUG.

VC-3
VC-3
5

VC-12
VC-12
VC-12
6
The TUG shown in figure 20 has 3 pointers for 7 Multiplex
Section
VC-12 Virtual Containers. The VC-12 is allocated 8 Overhead

a fixed space that is a little larger that it requires so 9

it has room to move around. Figure 20

The Pointer shows the location of the first byte of Bandwidth Efficiency
the VC-12.
N x 64 Kbit/s VC-12 (2.3 Mbit/s)

So when the VC-12 arrives at it destination the 2.048 Mbit/s VC-12 (2.3 Mbit/s)
receiving device will locate the VC-12 by looking at 8.192 Mbit/s VC-3 (48.9 Mbit/s)
the pointers.
10 Mbit/s VC-3 (48.9 Mbit/s)

3.9 Bandwidth Efficiency 34 Mbit/s VC-3 (48.9 Mbit/s)

Figure 21 shows a number of common data rates 45 Mbit/s VC-3 (48.9 Mbit/s)
and the corresponding Virtual Containers that would
need to be used to carry these data streams. 100 Mbit/s VC-4 (150 Mbit/s)
Figure 21

4. Conclusion
SDH has addressed the weaknesses of PDH. It transmits data in Virtual Containers and uses pointers to locate a
low speed channel in a high speed trunk.

Carriers like SDH because it provides:-

A Robust Ring Architecture with Self Healing Capabilities

Good Provisioning and Management Attributes
Strong International Standards