Traffic Generation of IEC 61850 Sampled Values

Jakub W. Konka , Colin M. Arthur , Francisco J. Garcia and Robert C. Atkinson

Dept.

of Electronic and Electrical Engineering

University of Strathclyde
Glasgow G1 1XW
Email: [email protected], [email protected]
Measurement Research Laboratory
Agilent Technologies
Edinburgh EH12 9DJ
Email: {colin arthur, frankie garcia}@agilent.com

AbstractThe work presented in this paper is targeted at

the first phase of the test and measurements product life cycle,
namely standardisation. During this initial phase of any product,
the emphasis is on the development of standards that support new
technologies while leaving the scope of implementations as open
as possible. To allow the engineer to freely create and invent
tools that can quickly help him simulate or emulate his ideas
are paramount. Within this scope, a traffic generation system
has been developed for IEC 61850 Sampled Values which will
help in the evaluation of the data models, data acquisition, data
fusion, data integration and data distribution between the various
devices and components that use this complex set of evolving
standards in Smart Grid systems.

I. I NTRODUCTION
The utility industry across the globe has been investigating possible ways of addressing grand challenges such as
generation diversification, more effective electricity demand
response, more efficient energy conservation, and reduction
of the industrys overall carbon footprint [1]. As one viable
solution the concept of a Smart Grid has been proposed.
A Smart Grid can be defined as a unified, fully interoperable,
communications-enabled electrical system which aims at revolutionising traditional power systems through the introduction
of condition monitoring and intelligence.
Many of the existing standards are being updated and new
standards are being developed incrementally in order to ensure
a timely and smooth transition from contemporary electrical
transmission systems to Smart Grids [2]. One such example
is the international standard IEC 61850. Originally developed
to define communications within electrical substations, it has
recently been extended to cover Smart Grid relevant aspects,
such as communications between substations, hydro power
plants and distributed energy resources (DERs) [3].
IEC 61850 standardises the way devices at the substation
level (and beyond) should communicate with each other to
achieve full interoperability. This is crucial for Smart Grids
as in such a complex system there is no room for possible
communication breakdowns between equipment provided by
different vendors. Therefore, thorough understanding of the
standard, and existence of realistic and accurate simulation
models of the protocols specified therein is a necessity.

Furthermore, as intelligent energy management evolves, like

with the Internet and Cyberspace, tussle will exist due to
the many different stakeholders [4]. In the case of intelligent
energy management the stake holders will include transmission
system operators, distribution network operators, equipment
manufacturers, regulators, ICT service providers, governments,
etc. Where tussle exists, it becomes imperative to ensure
that rigorous validation, verification and conformance testing
processes are put in place to ensure interoperability and
facilitate open competition in a thriving market place.
The test and measurement product life cycle can be broadly
categorised into standardisation, certification, interoperability
(IOT) field trials, monitoring and optimisation. For a test
and measurement organisation, the goal is to deliver a set
of test and measurement tools that can speed up the process
of any product moving through this cycle to get it into the
market place as fast as possible. The tools developed will
help any organisation take their product from research and
development, through validation, manufacturing, deployment
and future evolution.
The work presented in this paper is targeted at the first
phase of the test and measurements product life cycle, namely
standardisation. During this initial phase of any product, the
emphasis is on the development of standards that support new
technologies while leaving the scope of implementations as
open as possible. To allow the engineer to freely create and
invent tools that can quickly help him simulate or emulate his
ideas are paramount. Within this scope, a traffic generation
system has been developed for IEC 61850 Sampled Values
which will help in the evaluation of the data models, data
acquisition, data fusion, data integration and data distribution
between the various devices and components that use this
complex set of evolving standards in Smart Grid systems.
The paper is structured as follows. In Section II a brief
overview of IEC 61850 standard is given. Section III summarises simulation modelling research of the topic conducted
by other researchers. Section IV describes building blocks of
the Sampled Values traffic generator, and Section V presents
the steps taken and results of the validation process. Finally,
Section VI discusses future work and Section VII draws
conclusions.

Other substation
devices (e.g. HMI)

Time-critical
OSI stack
Sampled
Values

GOOSE

Direct
mapping

GSSE

GSSE
Specific

Time Sync

Client-Server
MMS

Application

UDP

TCP

Transport

IP

HMI Human-Machine Interface

P&C Protection & Control
MU Merging Unit
CB Circuit Breaker

Network

Ethernet (ISO/IEC 8802-3)

Data Link

Physical Medium

Physical

Gateway to other
substations
Switched
Ethernet

Station bus

P&C

Switched
Ethernet
MU

Process bus
CB
Power
line

Fig. 1.

Communication services defined in IEC 61850

Fig. 2.

Simplified protection & control system used within a substation

II. OVERVIEW OF IEC 61850

The first edition of the IEC 61850 standard comprises 12
parts in total. Each part describes different aspects of the
communications taking place within a substation automation
system. For example, part 7 presents the logical view of the
communication system, i.e., logical models of devices, abstract
communication service interface (ACSI), etc., while parts 8
and 9 specify how these logical concepts ought to be mapped
into a specific communications infrastructure, i.e., protocol
stack.
Throughout this investigation part 9-2 of the standard was
scrutinised in-depth since it describes how Sampled Values
protocol should be designed and mapped to Ethernet (ISO/IEC
8802-3). Part 9-1 was also analysed, however, in less detail
since it is likely to be withdrawn from the future releases of
the standard [5].
IEC 61850 defines 5 types of communication services as
shown in Fig. 1. The first three types are time-critical and
are used in protection and control schemes of the substation.
They include Sampled Values (SV) and Generic Object Oriented Substation Event (GOOSE) protocols which are mapped
directly to Data Link layer for reduced protocol overhead and
hence increased performance; and Generic Substation State
Event (GSSE) protocol which features its own custom protocol
mapping. The remaining two types of services, i.e., Time
Sync and Client-Server Manufacturing Message Specification
(MMS), deal with time synchronisation and management of
the substation devices respectively.
In order to better understand how these services are used
within a substation consider the simplified protection & control
system shown in Fig. 2. A Merging Unit (MU) accepts multiple analogue current and voltage samples coming from current
and voltage transformers which are connected directly to the
power line. It then acts as an analogue-to-digital-converter
converting the samples into digital format and encapsulating
them along with meta-data to put these measurements in
context as SV packets, and sends them over the switched
Ethernet network to Protection & Control (P&C) units [6].
The P&C extracts the samples, and examines them to detect
evidence of a fault in the protected zone of the network.
If the outcome of the examination is positive, it issues a
Trip message encapsulated as GOOSE/GSSE packet to the
Circuit Breaker (CB) unit which then trips the appropriate
switchgear and isolates the fault [7]. Afterwards, CB advises

other monitoring units of the action by issuing GOOSE/GSSE

packet. The exchange of time-critical messages described
above is based on the publisher/subscriber messaging model
in which one or more subscribers subscribe to the publisher
stating that they want to be notified of a particular event (in
this case, to act as sinks for SV or GOOSE/GSSE packets).
Time Sync service which was not mentioned in the example
above is substation-wide in scope and is used to distribute
common timing information across substation devices which
require such information. For instance, MUs in order to
correctly handle processing of current and voltage samples
need to be provided with timing information. The two most
popular methods for time synchronisation in the industry
include Global Positioning System (GPS) and the more recent
IEEE 1588 Precision Time Protocol (PTP).
Client-Server MMS, similarly to Time Sync service, is
substation-wide in scope and is used to manage substation devices attached to both process and station buses. In operation,
it is quite similar to Simple Network Management Protocol
(SNMP), and is also based on the client/server messaging
model in which the client polls the server for specific data.
The criticality of rigorous assessment and appraisal of interoperability is inherent to the widespread adoption of this standard within an industry that is risk averse and safety critical.
A facility to validate protection and automation functionality
within an increasingly integrated ICT system supporting the
Smart Grid is essential to its realisation. This work supports
this ambition.
III. L ITERATURE OVERVIEW
Several research investigations looked into the topic of
simulation modelling of communication services defined in the
IEC 61850 standard. Sidhu and Yin produced two publications
[8], [9] showing how the OPNET Modeler network simulator
can be used to model the substation communication network
and analyse the networks dynamic performance (i.e., packet
delay characteristics of time-critical services). Their model of
SV traffic generator mapped packets directly onto Ethernet
and, in their latter publication, included handling of IEEE
802.1Q VLAN tags. Moreover, in [9], the model additionally
appended SV-specific header, application service data units
(ASDUs), and application protocol data unit (APDU) according to IEC 61850-9-1.

SV Publisher

Application Layer

Type Length

SV Subscriber

savPdu

Packet flow

VLAN Interface

Data Link Layer

FDEthernet
FDChannel
Generator
Node

Fig. 3.

VLAN Interface

Switched
Ethernet

Value
Type Length

noASDU

0x80

Sequence of ASDU

0xA2 Length

Sequence ASDU 1
svID

FDEthernet
Physical Layer

0x60 Length

Value

10
0x30 Length

ASDU 1
0x80

10-34

Value

smpCnt

0x82

Value

confRev

0x83

smpSynch

0x85

Value

Sequence of Samples

0x87

64

Samples
(each
8-bytes
long)

FDChannel
Sink
Node

OSI-7 reference model of the SV traffic generator

Kanabar et al. [10] presented results of modelling communication networks for distributed automation systems with DERs
in the OPNET Modeler. In their investigation, SV protocol was
mapped directly onto Ethernet layer, and included VLAN tags.
However, no information was provided whether SV-specific
header, ASDUs and APDU were also appended. Similarly,
Ali and Thomas [11] did not specify whether any SV-specific
fields were attached to the packet while modelling various
communication scenarios within a substation in the OPNET
Modeler.
Finally, Liang and Campbell [12] concentrated on implementing and simulating ACSI in J-Sim open-source simulator.
Their goal was to explore security aspects of the standard and
in particular of IEC 61850-7.
IV. D ESCRIPTION OF THE T RAFFIC G ENERATOR
The model of SV traffic generator developed during this
investigation is based on part 9-2 of the IEC 61850 standard
[13], [14], and on IEC 61850-9-2 LE implementation guidelines [15]. It was implemented in NS3 open-source network
simulator, version 3.8 [16]. NS3 was chosen out of many
other available simulators for two reasons: 1) it is opensource, hence it is possible to alter the core of the simulator if
necessary; and 2) it features uniform coding structure, i.e., both
simulation scripts and its core are written in C++ programming
language, hence there is no need to create language bindings
as is the case in NS2 for example (OTcl/C++ bindings).
Fig. 3 depicts a conceptual model of the traffic generator
in the spirit of OSI-7 reference model. The generator node
is responsible for generating packets carrying SV protocol
(or simply SV packets) while the sink node acts as their
addressee. Both nodes implement 3 out of 7 OSI layers as
required by [14]: application, data link, and physical layer.
The functionality of the application layer is implemented by
SV Publisher application in the generator node, and by SV
Subscriber application in the sink node. The functionality of
the data link layer is split into two parts: VLAN Interface and
FDEthernet sublayers; while the functionality of the physical
layer is implemented by FDChannel. VLAN Interface and
FDEthernet sublayers together with FDChannel model the
behaviour of full-duplex Ethernet with IEEE 802.1Q support,

Fig. 4.

Sequence ASDU 2

0x30 Length

ASDU 2

Sequence ASDU 10

0x30 Length

ASDU 10

Structure of modelled SV APDU/ASDU (ASN.1/BER TLV triplets)

i.e., payload coming from upper layers is encapsulated inside

the Ethernet-II (EtherType) type framing with VLAN tagging.
All parts of the traffic generator were fully designed and
engineered in the simulator in order to gain first-hand experience in working with the standard. Moreover, since the main
aim was to create a realistic traffic generator, as packets go
up/down the protocol stack appropriate headers and trailers are
appended/removed on a layer basis similarly to show how it
occurs in a real communication scenario.
A. SV Publisher/Subscriber Application
Every SV packet is created by SV Publisher application, and
upon creation consists of a series of SV ASDUs encapsulated
inside the SV APDU. The byte structure of SV APDU and
SV ASDU implemented in the model is based on the Abstract
Syntax Notation One (ASN.1) and Basic Encoding Rules
(BER) Type-Length-Value (TLV) triplets specified in [14].
The structure is shown in Fig. 4. At present, the model of
SV APDU/ASDU does not include any of the optional TLV
triplets, i.e., datSet, refrTm, smpRate and smpMod. Moreover,
the majority of the implemented triplets have their Value fields
set to 0 or an arbitrary value to simply preserve the structure of SV APDU/ASDU rather than carry any meaningful
information. Future work will involve adding optional triplets,
and setting Value fields to something more meaningful as the
SV traffic generator will be improved, and models of other
protocols such as GOOSE/GSSE will be added.
SV-specific header is appended to SV APDU before it is
sent down the protocol stack to lower layers. The header
is 8 bytes long and comprises 4 fields: APPID, Length,
Resrv1, and Resrv2 (see Fig. 5, topmost packet structure).
The APPID field stands for Application Identifier [14], and
is used to select those Ethernet frames which contain SV
APDU, and to distinguish between GOOSE/GSSE and SV
protocols. APPID is always 2 bytes long and in case of SV
protocol has the two most significant bits (MSBs) set to 01,
while the remaining 14 bits are the actual ID used within the
substation communication network. Therefore, the values of
APPID range from 0x4000 to 0x7FFF. By default, it should
be set to 0x4000.
The Length field is also 2 bytes long, and indicates the
length of the packet including the SV-specific header, i.e., if

APPID

Length=
m+8

2
Resrv1

Fig. 5.

6
MAC
Src

Resrv2

46-(m+8) bytes

APDU

Padding

Application
layer

VLAN
TCI

MAC
Dst

EtherType=
Payload
0x88BA

2
VLAN TPID=
Payload
0x8100

VLAN Interface
sublayer

4
FCS

FDEthernet
sublayer

Encapsulation of SV APDU as it goes down the protocol stack

the length of the APDU is m bytes and since the length of

the SV-specific header is 8 bytes, Length field shall be set to
m + 8.
The Resrv1 field stands for Reserved 1. It is 2 bytes long,
and features a set of blocks of bits each serving different
function. The MSB, if set to 1, indicates that this SV packet
has been issued by a test device. The following 3 bits are
reserved for future standardised application, and by default,
are set to 0. The remaining 12 bits are reserved for security
purposes, and by default, should also be set to 0x000. The
Resrv2 field stands for Reserved 2. It is 2 bytes long, and is
used to denote additional security parameters carried by this
SV packet. By default, it should be set to 0x0000.
In our model, APPID and Resrv2 fields assume their default
values, that is APPID is set to 0x4000 and Resrv2 to 0x0000.
Resrv1 field is set to 0x8000 to indicate that packets were generated by a test device. Length field is updated automatically
based on the actual length of SV APDU. If the total length of
the packet (SV APDU + SV-specific header) is smaller than 46
bytes, SV Publisher application will pad it with zeros in order
to ensure that it conforms with the requirements of Ethernet,
i.e., the minimum length of the payload encapsulated inside
the Ethernet frame must not be smaller than 46 bytes.
The SV Subscriber application on the other hand acts
as a simple sink for SV packets. It basically removes SVspecific header and padding (if any) from a packet which was
forwarded up to the application by VLAN Interface sublayer.
B. VLAN Interface Sublayer
A VLAN Interface sublayer was designed to imitate the
behaviour of an IEEE 802.1Q VLAN-aware end station/device,
i.e., a network node which is capable of explicitly VLAN
tagging the frame before it is ever sent out [17]. It is clearly
stated in [14] that frames carrying SV APDU have to be
VLAN tagged explicitly. The main task of this sublayer is
to append/remove VLAN Tag Control Information (TCI) and
EtherType field to/from the packet received from upper/lower
layer (see Fig. 5, middle packet structure). The VLAN TCI
comprises three fields [17]: Priority, Canonical format indicator (CFI), and VLAN ID.
The Priority field is 3 bits long, and is used to denote the
priority which the frame should be treated with when travelling
through priority-aware switched Ethernet network. Its value

can range from 0 to 7 according to IEEE 802.1p. There are

no strict rules as to how those values should be distributed
among different types of traffic but, by default, for SV-type
payload, priority should be set to 4 [14].
The CFI field is one bit long, and in general, refers to the bit
ordering format of the bytes within a frame. Its precise meaning however depends on the underlying technology (Ethernet,
Token Ring, etc.). In Ethernet networks, frames are normally
sent with CFI bit set to 0.
The VLAN ID field is 12 bits long, and is used to explicitly
identify frames VLAN association. For SV-type payload, if no
VLAN segments were configured, it should be set to 0x000
[14]. Otherwise, since SV protocol needs to have potentially its
own bandwidth allocation, it should be set to a value different
from the one used by GOOSE/GSSE protocols.
The value of each of the VLAN TCI composite fields is
determined at the application layer, hence in this case by
SV Publisher application, and is passed to VLAN Interface
sublayer together with the pointer to SV packet and the value
of EtherType field. The model assumes the following default
values: Priority set to 4, CFI to 0, and VLAN ID to 0x000. The
EtherType field which uniquely identifies the payload/protocol
carried within the Ethernet frame is set to SV Protocol ID, that
is 0x88BA.
When the packet is received from the lower layer, the VLAN
Interface sublayer removes VLAN TCI and EtherType field.
It then inspects the removed EtherType field and compares
it with the SV Protocol ID. If the two match, the packet is
forwarded up the protocol stack to SV Subscriber application.
Otherwise, the packet is dropped at this sublayer. This mechanism was developed to ensure that only VLAN tagged frames
carrying SV-type payload are forwarded up to SV Subscriber
application.
C. FDEthernet Sublayer
The FDEthernet sublayer was designed to imitate the
behaviour of full-duplex Ethernet network interface card.
Therefore, its main functions include: enqueueing/dequeueing
frames for transmission over FDChannel; appending/removing
destination and source MAC address fields, and VLAN Tag
Protocol Identifier (TPID) field to/from the frame going
up/down the protocol stack (see Fig. 5, lowermost packet
structure); and calculating and appending/removing the Frame
Check Sequence (FCS).
The value of VLAN TPID (i.e., VLAN Protocol ID,
0x8100) is provided to FDEthernet sublayer by VLAN Interface together with the pointer to payload. When packet
is received from lower layer, i.e., FDChannel, and passes
checksum control, FDEthernet sublayer forwards the frame
up the protocol stack only if the third field of the EthernetII header (i.e., EtherType field if no VLAN tagging is employed) matches the VLAN Protocol ID. Otherwise, the packet
is discarded at this sublayer. This mechanism ensures that
only VLAN tagged frames are forwarded to VLAN Interface
sublayer.

Moreover, since frames carrying SV-type payload are

VLAN tagged, the Ethernet maximum transmission unit
(MTU) needs to be increased from 1,518 to 1,522 bytes
according to IEEE 802.3ac supplement [17]. MTU is used to
determine maximum frame length (Ethernet-II header inclusive) at the Ethernet layer which if exceeded causes frames to
be fragmented. In our model, MTU is executed by FDEthernet
sublayer.
D. FDChannel
FDChannel acts as a physical transmission medium, and can
be configured to behave like any type of full-duplex physical
medium, e.g. twisted pair or fibre link.
V. M ODEL VALIDATION
In order to validate the model of SV traffic generator, a
simple simulation scenario was created. Packet traces gathered
from the simulation were saved in PCAP file format which
is natively supported by NS3 [16], and can be analysed in
Wireshark network protocol analyser [18]. If the structure of
generated SV packets is dissected by Wireshark correctly, we
can conclude that the traffic generator behaves as required.
A. Simulation Scenario
The topology of the simulated network consisted of 4 nodes:
2 SV Publisher end devices (node0 and node1) were generating SV packets and sending them to SV Subscriber end device
(node2) through a priority-aware L2 Ethernet switch. All links
were modelled as full-duplex with capacity of 100 Mbps and
propagation delay of 1 ms.
Each generated SV packet was 997 bytes long including all
the headers, i.e., SV-specific header, VLAN tags and EthernetII header (1001 bytes including FCS). Each SV APDU was
971 bytes long and comprised 10 SV ASDUs, each 96 bytes
long. Moreover, each SV ASDU comprised: buffer worth of
8 sampled values (64 bytes long) padded with zeros; Value
field of svID TLV triplet set to NS3Generated (13 bytes
long including null character); Value fields of confRev and
smpSynch TLV triplets set to 1; and all other Value fields of the
remaining (non-optional) TLV triplets set to 0. Furthermore,
node0 generated SV packets with the total bit rate of 80 Mbps,
while node1 with the total bit rate of 70 Mbps.
B. Simulation Results
Packet traces were gathered at the SV Subscriber node
(node2). Furthermore, in all of the traces, SV nodes assume the following MAC addresses: 00:00:00:00:00:01 for
node0, 00:00:00:00:00:03 for node1, and 00:00:00:00:00:05
for node2.
Traces were analysed in the latest stable release of Wireshark (at the time of writing; version 1.4.0) [18], on Ubuntu
Linux 10.04.
Fig. 6a shows the result of dissecting SV packets by
Wireshark. The length of the SV packet including all the
headers, i.e., SV-specific header, VLAN tags and EthernetII header, is displayed to be 997 bytes as desired. Moreover,

SV-specific header has its fields set in accordance with the

assumptions of the modelling process: APPID is set to 0x4000,
Resrv1 to 0x8000 and Resrv2 to 0x0000. Length field, which is
supposed to indicate the actual length of SV APDU including
the length of SV-specific header, is set to 979. It agrees with
the assumptions stated earlier and is also proven to be the case
by Wireshark the bottom of Fig. 6a shows that Wireshark
interpreted the length of SV APDU including SV-specific
header to be 979 bytes as well. SV APDU is shown to consist
of 10 SV ASDUs as desired. Furthermore, Value fields of
TLV triplets in each SV ASDU also correspond to the assumed
values, for example svID is set to NS3Generated as assumed
in Section V-A.
Finally, Fig. 6b depicts the contents of the buffer which
carries sampled values of current and voltage. Although in our
model the buffer is padded with zeros throughout, its structure
is intact and can be populated with real values in the future.
All in all, this proves that our SV traffic generator successfully
models the structure of SV packets accurately.
VI. F UTURE W ORK
There are a number of potentially fruitful research issues
worthy of further investigation. Firstly, the model of SV traffic
generator should be improved by feeding SV APDU/ASDU
with realistic data. The more realistic the traffic generator,
the more accurate the output from the simulator and hence,
the greater the insight into different issues of communications
within the substation (and beyond). Traffic with realistic data
allows to better scale the network still in the simulation
modelling stage; thus, hastening the process of installing the
communications network as well as making it more reliable
from the very beginning.
Secondly, the model of SV traffic generator described in
this paper could be used as a building block for GOOSE
and GSSE traffic generators. Implementation of these two
protocols is the next step in understanding the IEC 61850
standard thoroughly and creating a realistic simulator of the
substation communication network based on this standard.
Lastly, models of Time Sync and Client/Server MMS
services should be added and tied up with models of
GOOSE/GSSE and SV protocols. This would pave the way
for more Smart Grid relevant simulation research such as
modelling of communications between substations.
VII. C ONCLUSIONS
Smart Grids are a major leap forward in the functionality
of electrical systems. As such, this new emerging technology
requires advanced and bullet-proof communications on all
levels of abstraction. The international standard IEC 61850 is
the most likely candidate for standardisation of communications on substation level and beyond. Therefore, its thorough
understanding and existence of accurate and realistic models
of the protocols and communication services described therein
is a necessity.
In this paper, an accurate and realistic model of Sampled
Values traffic generator was described. Packet traces generated

(a) SV-specific header and SV APDU/ASDU

Fig. 6.

(b) Buffer containing sampled values

Structure of SV packets generated during the simulation dissected by Wireshark

using the model were shown to be realistic as they included

the full byte structure of Sampled Values APDU and ASDU.
Moreover, they were also correct since Wireshark network protocol analyser dissected them without any glitches whatsoever.
The model can be used to assess all critical to safety
first utility applications such as: the interoperability with
realistic traffic levels; the assessment of bandwidth and potential erroneous operation; resilience to interference/noise;
etc. Furthermore, the development of the model from first
principles during this investigation proved to be very useful
and gave great insight into how communications is managed
in the IEC 61850-enabled substation.
ACKNOWLEDGMENTS
The authors would like to thank Peter Cain from Agilent
Technologies for providing access to an up-to-date report from
NIST proceedings on the development and standardisation of
Smart Grids, and Paul Crolla, Dr. James Irvine and Professor
Graeme Burt from University of Strathclyde for their valuable
comments on the contents of this paper.
R EFERENCES
[1] H. Farhangi, The Path of the Smart Grid, in IEEE Power and Energy
Magazine, vol. 8, pp. 18-28, Jan.-Feb. 2010.
[2] NIST, NIST Framework and Roadmap for Smart Grid Interoperability
Standards, Release 1.0. NIST Special Publication 1108. Office of the
National Coordinator for Smart Grid Interoperability, January 2010.
[3] International Standard IEC 61850: Communication networks and systems for power utility automation. IEC, 2002-10.
[4] D. D. Clark and J. Wroclawski, Tussle in Cyberspace: Defining
Tomorrows Internet, in SIGCOMM02, 19-23 August 2002, Pittsburgh,
Pennsylvania, USA.
[5] K. Schwarz, What is Edition 1 and Edition 2 of IEC 61850?
News on IEC 61850 and Related Standards, Mar. 3, 2010.
[Online].
Available:
http://iec61850-news.blogspot.com/2010/03/
what-is-edition-1-and-edition-2-of-iec.html. [Accessed: Aug. 23,
2010].

[6] M. C. Janssen and A. Apostolov, IEC 61850 Impact on Substation

Design, in Transmission and Distribution Conference and Exposition
2008. T&D. IEEE PES, pp. 1-7, April 2008.
[7] H. Ito and K. Ohashi, High Performance IEC 61850 GOOSE and
Protection Relay Testing, in PAC World Magazine, Winter 2008
Issue. [Online]. Available: http://www.pacw.org/no-cache/issue/winter
2008 issue/protection goose/high performance iec 61850 goose
and protection relay testing.html. [Accessed: Aug. 23, 2010].
[8] T. S. Sidhu and Y. Yin, IED Modelling for IEC61850 Based Substation
Automation System Performance Simulation, in Power Engineering
Society General Meeting, 2006. IEEE, pp. 1-7, 18-22 June 2006.
[9] T. S. Sidhu and Y. Yin, Modelling and Simulation for Performance
Evaluation of IEC61850-Based Substation Communication Systems, in
IEEE Transactions on Power Delivery, vol. 22, pp. 1482-1489, July
2007.
[10] P. M. Kanabar, M. G. Kanabar, W. El-Khattam, T. S. Sidhu and
A. Shami, Evaluation of Communication Technologies for IEC 61850
based Distribution Automation System with Distributed Energy Resources, in Power & Energy Society General Meeting, 2009. PES 09.
IEEE, pp. 1-8, 26-30 July 2009.
[11] I. Ali and M. S. Thomas, Substation Communication Networks Architecture, in Power System Technology and IEEE Power India Conference,
2008. POWERCON 2008. Joint International Conference on, pp. 1-8,
12-15 Oct. 2008.
[12] Y. Liang and R. H. Campbell, Understanding and Simulating the IEC
61850 Standard, 2008. [Online]. Available: https://www.ideals.illinois.
edu/handle/2142/11457. [Accessed: Aug. 23, 2010].
[13] International Standard IEC 61850-9-2: Communication networks and
systems for power utility automation Part 9-2: Specific Communication
Service Mapping (SCSM) Sampled values over ISO/IEC 8802-3,
1st ed. IEC, 2004.
[14] International Standard IEC 61850-9-2: Communication networks and
systems for power utility automation Part 9-2: Specific Communication
Service Mapping (SCSM) Sampled values over ISO/IEC 8802-3,
2nd ed. Draft for public comment. IEC, 2009.
[15] UCA International Users Group, IEC 61850-9-2 LE: Implementation
Guideline for Digital Interface to Instrument Transformers Using IEC
61850-9-2. UCA, 2004.
[16] NS3 Open-Source Network Simulator, version 3.8. [Online]. Available:
http://www.nsnam.org. [Accessed: Aug. 23, 2010].
[17] R. Seifert and J. Edwards, The All-New Switch Book: The Complete
Guide to LAN Switching Technology, 2nd ed. Indianapolis, IN: Wiley
Publishing, 2008.
[18] Wireshark Network Protocol Analyzer, version 1.4.0. [Online]. Available:
http://www.wireshark.org. [Accessed: Aug. 30, 2010].