Background:Topology may be used as a term to

describe geometric characteristics of the interconnection of points in two or

three dimensional space. Network may be used as a term to describe the
existence of a connection between various points, typically for the purpose of
exchanging information. When these terms are combined, as in network
topology, the implication is to describe the geometric shape of the lines of
communication, such as ring or star.
This concept has been understood for
centuries, but has only become critical at the device level within the past few
decades. In the earliest days of programmable logic controllers (PLC) devices
were connected to the control environment in a method called point-to-point
wiring. Every input sensing device or output actuating device was individually
connected to the PLC, controller, or data acquisition instrument using point-topoint hard-wired connections. This was one of the most expensive aspects of
automation, because it is so labor intensive and requires so much material
infrastructure, such as like wire and conduit.
When PLCs were introduced to
automation in the 1970s they included network communication capabilities
which allowed them to pass information between themselves and to central
control systems. The earliest PLCs were the Allen-Bradley and Modicon
systems, which included their respective networks named Data Highway and
Modbus, both of which still exist today, although highly evolved to be used over
Ethernet. But, in their earliest form they used adaptations of serial RS-485
communication to accomplish a relatively flexible and high speed multi-drop
communication network for that era.

Besides communication, PLCs

introduced several other changes to the automation paradigm which would
change the topology of automation forever. The biggest change was the
introduction of the Moores Law era. Moore was not the first to make his
prediction, but in 1965 he made it popular by publishing a quantified rate in a
magazine article. Moore basically stated that the size of electronic circuitry
would shrink very quickly while its functional capacity would increase very
rapidly. The PLC is one piece of evidence for this evolution, as it replaced large,
expensive, heavy mechanical relays with much smaller electronic solutions.
Since then this trend has not slowed in development, influencing not only the
size and processing power of PLCs, but the communication capability within
the automation paradigm, and the devices connected to the PLCs.
One aspect of electronic design in networked
systems with respect to Moores Law is to move these more compact and more
powerful electronics closer to the end nodes of the network. This is also
facilitated by faster communication technology. These end nodes are the
various device level components; such as sensors and actuators, that are
typically selected first in the component selection of the design phase. From the
transducer layer to data collection and concentration layer is a digital
communication link that is almost always a fieldbus network layer. Topology
gives us a way of visualizing this network and its interconnections.
There are three important concepts at work
here. First, the network, thats the transport layer for the data. It is composed
of nodes. These nodes can be end-points, like devices, connectors, like hubs
and switches, conversion nodes, like protocol shifters or gateways. The next is
the media that connect the nodes. The media can be wired or wireless. Third, is
the protocol. The protocol is the encoding that permits the data to move across

the media to or from the nodes. Topology refers to both the physical network
(wires, hubs, switches, devices, etc.) and the logical networkwith all its
interconnections.
Device:Some of the most rapid and radical technology change is occurring at the
device level. The device level may be called edge technology because it is at the
boundary dividing the real world environment from the sensing and actuating
done by technology. Devices may also be considered elemental, because they
are the smallest finite point of interface of technology to the real world.
For instance, a thermistor may be considered a
device. A thermistor is a special kind of resistor, a small electronic component,
typically two-wire, very inexpensive to manufacture, yet very effective and
reliable as a temperature sensing device. The sensing characteristic of a
thermistor is that its resistance changes very predictably and consistently with
temperature in the range for which it is designed. In many applications a
thermistor is the only component in a wall mounted temperature sensing
device. These devices may be seen on the walls and hidden in the decor of
many large commercial spaces as part of the heating and cooling system. These
are some of the simple end nodes or edge devices between the environment
being sensed and controlled, and the technology that provides this managed
comfort.
At the elemental level, these devices always fall
into one of two general categories; sensors or actuators. These two types of
devices give people a view into the environment for which they have a concern
that is, sensing, and offer the ability to influence or make some adjustment to
that environment, that is, actuating. When sensing and actuating occurs
automatically, it is called automation and requires some control mechanism to

manage the actuation based on some limits of the parameters being sensed.
When the control component of an automated paradigm is electronic, it is often
called the control device. However, it is not the same as sensing and actuating
devices. Control devices are usually significantly more complex than edge
devices, and in the past, have physically been distant from the real world edge.
For the control device to. 3 communicate with edge device, a communication
network was necessary, as well as a device to transduce the raw edge device
signal, sometimes called the field signal to a digital electronic signal the control
device could receive, transmit, and process. Normally, in an automation
paradigm, the edge device level is specified first. The real world environment to
be sensed and perhaps automatically controlled is assessed for the type of
sensing necessary and devices are selected with a mind to how they will be
connected to the technology that will convert real world parameters to
electronic data to be stored, processed, perhaps for analysis, control, or
historical records. If control is prescribed, then points where physical action
must be taken to manage the environment are identified and appropriate
actuator devices are selected. Network Today most electronic communication is
done serially, even at computer bus levels, in part, because the transmission
speed is consistent with Moores Law, and has increased by orders of
magnitude. Ethernet, Universal Serial Bus (USB), Serial Advanced Technology
Attachment (SATA), Enhanced PCI (PCIe), FireWire, Thunderbolt, and Scalable
Link Interface (SLI), among many others, are all serial, and very high speed.
Component sizes have correspondingly shrunk making it possible to include
very high speed connectivity to almost any device. Speed and component size
will continue to be a challenge, speed will continue increase and components
will continue to shrink. The greater challenge is optimization of data transfer to
take full advantage of the speed. Technologies to increase efficiency, reliability,

determinism include physical media types, communication protocols, and

network topologies. Media There are many media over which weve run
networks. The configuration of the cabling layout is the physical topology of the
network. Starting with coaxial cable, and proceeding to multi-conductor twisted
pair and then to fiber optic cables, the choice of media determines the size,
distance and data throughput maximum of the network. Since the early 2000s,
network media include radio transmission over special purpose low power
radios like WiFi and 802.15.4. The logical topology of the network refers to the
way that the signals act without reference to the physical interconnection of the
network devices or nodes. This is not always the same as the physical topology
of the network. Logical topologies are about the path the data takes between
the nodes of the network, rather than the physical connections of the network
nodes themselves. The logical topologies are mostly determined by the network
protocols. Protocols Networks are not limited to Ethernet. Everything from
Modbus RTU over serial RS-485 multidrop, to HART, to PROFIBUS, to
Foundation Fieldbus to low power wireless personal area networks (WPAN) can
be configured in many of these topologies. Wired and wireless networks are
made up of different media (wires and radio), but use the same network
topologies, and can be mixed and matched in complicated ways. Just as wired
networks use different protocols, so do wireless networks. There are at least five
commonly used wireless network protocols that are standards-based: Zigbee,
IEC62591-WirelessHART, ISA100.11a

and

the Chinese WIA-PA

wireless

standards use basically similar firmware riding on top of IEEE 802.15.4 low
power mesh radios. IEEE 802.11xx, commonly called WiFi, uses a different set
of radios but can be designed in similar topologies. Zigbee, Wireless HART,
ISA100 and WIA-PA are incompatible and cannot be used in the same network.
There are numerous wireless sensor network protocols that are not standards

based. Surveys by Control magazine among others indicate that nearly 30% of
all wireless field sensor networks are using these proprietary protocols, and
that number does not appear to be declining.

Topologies
There are a finite number of ways nodes on a network can be connected, both
to the network and to each other. The basic topologies network engineers
observe in industrial automation are;
Point-to-point (or Line)
Multi-drop (or Bus)

Ring
Star
Tree
Mesh
Fully Connected Substation n

It is extremely important to remember that these topologies can be hardwired

or wireless connection paths. They are independent from the protocols and the
data that use these protocols over the network.
Point-to-Point
Point to point is the simplest topologytwo points on the network permanently
connected. There are two kinds of point to point connectivity. The first is a
dedicated permanent connection, a channel that is permanently associated
with two endpoints. A leased pair line from a SCADA RTU to a central head end
is a classic example of this type of topology.
The second type is switched. This type sets up a point to point connection on
demand wherever it is required. Analog telephony is an example of this, and
digital packet-switching technologies are logical extensions of this point to
point connectivity.
Multi-drop
In a bus, such as a fieldbus, all the devices are connected to a single bus cable.
The signal travels the full length of the bus until it finds the device with which
it is intended to communicate. If the device it encounters does not match the
address of the device it is seeking, the data continues on the bus until it finds
a device that does not ignore the data. Industrially, PROFIBUS and Foundation
Fieldbus are common fully digital fieldbus protocols that use bus topology. Bus
topologies are very inexpensive to implement and are therefore used often in
industrial automation for sensor level or device level network topologies.

The problem with bus networks is that they have a relatively high cost to
manage, and they can be a single point of failure. They are therefore used for
simple, short distance networks like machine control, or sensor networks.

Advantages

Easy to connect a computer or peripheral to a linear bus

Requires less cable length than a star topology

It works well for small networks.

Disadvantages

Entire network shuts down if there is a break in the main cable

Terminators are required at both ends of the backbone cable

Difficult to identify the problem if the entire network shuts down

Not meant to be used as a stand-alone solution in a large building

It is slow when more devices are added into the network

If a main cable is damaged then network will fail or be split into two
networks

Ring
A ring network can be thought of as a line network that connects its last node
back to its firstcreating a ring. Data travels in only one direction on the ring,
and each node acts as a repeater to retransmit the data at a high signal level.
So there is a transmitter and receiver in each node. Every node is therefore a
critical link. If one of the links in the ring is broken, data cannot travel around
the ring and the topology is broken.

Advantages:

Very orderly network where every device has access to the token and the
opportunity to transmit

Performs better than a bus topology under heavy network load

Does not require a central node to manage the connectivity between the
computers

Due to the point to point line configuration of devices with a device on

either side (each device is connected to its immediate neighbour), it is quite

easy to install and reconfigure since adding or removing a device requires

moving just two connections.
Point to point line configuration makes it easy to identify and isolate

faults.
Disadvantages
One malfunctioning workstation can create problems for the entire

network. This can be solved by using a dual ring or a switch that closes off
the break.

Moving, adding and changing the devices can affect the network

Communication delay is directly proportional to number of nodes in the

network

Bandwidth is shared on all links between devices

More difficult to configure than a Star: node adjunction = Ring shutdown

and reconfiguration

Star
The point-to-point topology is really a special case of the star topologyone of
the direct connections from the central node to the endpoint. Every node in a
star is connected to a central node, which can be either a hub or switch. The
switch serves the peripherals as if the switch was the server and the peripheral
devices were the clients. Any traffic on the network must pass through the

central server point or switch. Stars are considered very easy to implement and
maintain.

Stars can be more than five pointed, too. It is easy to add another node to the
network by simply connecting it, either physically or logically, so that the data
from the new node passes through the hub of the star. It is even possible to
stack stars, with point-to-point connections from the hub of one star to the
hub of another star. In this way, the entire star becomes an arm of the next
star. This is often referred to as a daisy-chained set of stars, or a distributed
star topology.
This is the most significant disadvantage of the star topology: the hub or switch
at the center of the star is a single point of failure.
Nowadays many substation vendors apply star topology in a station bus
because of the following
Advantages:
Easy to pinpoint the location of a problem and isolate and solve it
Low electromagnetic interference
Easy steps for troubleshooting
Modification flexibility, easier to maintain
Failure of one node does not affect the network

 Very high speeds of data transfer

High maintainability
Multicast island can be built, traffic is reduced, high latency tolerance
Easy plug-and-play of power equipment
Single Star Topology
In star topology each station (peer) is connected directly to a common central
node or switch. Transmission occurs through switches between components. It
is the only topology that can operate with an unmanaged switch. At the same
time, star topology is less reliable because all IEDs are connected to a single
central Ethernet switch which is highly susceptible to environmental and EMI
conditions of the power substation. A failure in one switch can disrupt the
network. Because a large number of bridges in a big substation network need a
large number of wires to connect the network, the cost of building such a
network with star topology is relatively high.

Double Star Topology

In order to enhance substation network reliability and availability, double star
or redundant star topology is often implemented. Others apply logical double

star topology by using double Ethernet switches. For logical double star
topology, each critical device should have a double access interface (DAN,
Double Access Node) and be connected to two independent LANs. PRP (Parallel
Redundant Protocol) or RSTP (Rapid Spanning Tree Protocol) can be applied
with a double network infrastructure for full redundancy. Figure 2 shows a
double star topology with two RSTPs in a station bus.

Tree
A tree network topology is a hierarchical network of at least three levels that
consists of a root node, connected to a node or more than one node lower in the
hierarchy, with point to point links that can be visualized as branches of the
tree. The tree is symmetrical with every node connected to a specific fixed
number of nodes at the next lower level.
Tree topologies are very scalable and resemble hybrid networks. They are very
durable and cost effective for wired networks, but less so for wireless networks
because of the need for all the nodes on a branch to go through other nodes

before reaching the trunk. This adds potential latency in the case of a sensor
network, but is quite workable in an enterprise network topology.
Mesh
Mesh networks have been used in Industrial Automation since the early 1980s,
predating Ethernet networks entirely. In the last ten to fifteen years, they have
become much more important because almost all the international standard
wireless protocols use a mesh topology.
Mesh networks follow Reeds Law for the number of possible connections: n (n1)/2. But that only works if the mesh is fully connected. That is, all the nodes
in the mesh connect either physically or logically with every other node.
This provides the very high level of redundancy needed in sensor networks of
all kinds, because the data packet will find its way from the origin node to the
destination node using any of the paths of the mesh.
Remember, it is not only topology, but other considerations like ease of
maintenance, or in the case of battery powered modules, the simple lack of
availability of an external power source that dictates other aspects of device
selections such as transducers and sensors

Certainly wireless sensor networks (WSN) are not the only place for careful
consideration of connectivity topology. But, the benefit of mesh reliability may
be outweighed by other factors when wireless is not possible. For instance, the
cost of a wired fully connected mesh is exorbitant beyond a very few nodes.
Wireless networks can be scaled nearly infinitely, using daisy chains of mesh
networks.

Tree
A tree topology is essentially a combination of bus topology and star topology.
The nodes of bus topology are replaced with standalone star topology networks.
This results in both disadvantages of bus topology and advantages of star
topology.
For example, if the connection between two groups of networks is broken down
due to breaking of the connection on the central linear core, then those two
groups cannot communicate, much like nodes of a bus topology. However, the
star topology nodes will effectively communicate with each other.

It has a root node, intermediate nodes, and ultimate nodes. This structure is
arranged in a hierarchical form and any intermediate node can have any
number of the child nodes.

But the tree topology is practically impossible to construct, because the node in
the network is nothing, but the computing device can have maximum one or
two connections, so we cannot attach more than 2 child nodes to the
computing device (or parent node). There are many sub structures under tree
topology, but the most convenient is B-tree topology whereby finding errors is
relatively easy.
1. A network that is based upon the physical hierarchical topology must
have at least three levels in the hierarchy of the tree, since a network
with a central 'root' node and only one hierarchical level below it would
exhibit the physical topology of a star.
2. A network that is based upon the physical hierarchical topology and with
a branching factor of 1 would be classified as a physical linear topology.
3. The branching factor, f, is independent of the total number of nodes in
the network and, therefore, if the nodes in the network require ports for
connection to other nodes the total number of ports per node may be
kept low even though the total number of nodes is large; this makes
the effect of the cost of adding ports to each node totally dependent upon
the branching factor and may therefore be kept as low as required
without any effect upon the total number of nodes that are possible.

4. The total number of point-to-point links in a network that is based upon

the physical hierarchical topology will be one less than the total number
of nodes in the network.
5. If the nodes in a network that is based upon the physical hierarchical
topology are required to perform any processing upon the data that is
transmitted between nodes in the network, the nodes that are at higher
levels in the hierarchy will be required to perform more processing
operations on behalf of other nodes than the nodes that are lower in the
hierarchy. Such a type of network topology is very useful and highly
recommended.
Advantages

It is scalable. Secondary nodes allow more devices to be connected to a

central node.

Point to point connection of devices.

Having different levels of the network makes it more manageable hence

easier fault identification and isolation.

An example of this network could be cable TV technology. Other examples are

in dynamic tree based wireless networks for military, mining and otherwise
mobile applications.
Disadvantages

Maintenance of the network may be an issue when the network spans a

great area.

Since it is a variation of bus topology, if the backbone fails, the entire

network is crippled.

An example of this network could be cable TV technology. Other examples are

in dynamic tree based wireless networks for military, mining and otherwise
mobile

applications.

The Naval

Postgraduate

School,

Monterey

CA,

demonstrated such tree based wireless networks for border security. In a pilot
system, aerial cameras kept aloft by balloons relayed real time high resolution
video to ground personnel via a dynamic self-healing tree based network.

Hybrid Topology
The protocol and topology options certainly contribute to the increasing
potential for significant complexity in the network layer at the device layer.
Network topologies can be ganged or daisy chained, so long as the nodes are all
running the same protocol, or there is a protocol converter at the point at
which a new protocol is needed. As networks mix protocols, as well as media
with both wired and wireless networks in the same plant or factory, hybrid
network topologies are becoming much more common. In a typical factory or
process plant, there might be for example, a wired Profibus network connected
to an Ethernet switch which is also connected to a wireless mesh network
gateway, and to a Modbus TCP/IP network that is running over Ethernet.
Device Level Network Topology
There are at least three significant factors contributing to the effect of Moores
Law on the evolution of device level technology with respect to the functions of
transducing and network communication moving to the edge. Certainly the
increasing speed and power of the device technologies is one aspect, and the
shrinking size of these technologies allows them to be packaged at the sensor

and actuator level. The third factor is the cost effectiveness of an adequately
flexible combination of an edge device, such as a sensor or an actuator, small
enough and fast enough to be packaged with the edge device, rugged enough to
handle the environments where edge devices are frequently used in the real
world

environment,

and

powerful

enough

to

handle

the

high

end

communication and perhaps some control capabilities that current technology

demands.
Can an input and an output be packaged together, such that the input could
control the output, in a small package, with an easily configured control
environment, and powerful communication capabilities to pass a the higher
levels of data collection and concentration? Advantech is one of the companies
continually pursuing smaller, faster, and more powerful. It is recognized that
the real goal is a 12 mm proximity switch, with a rugged micro DIN Ethernet
connector, that speaks a popular protocol like Modbus/TCP, has sufficient
processing power to condition the field signal to engineering units and
transduce the signal to digital, and perhaps provide some level of simple remote
distributed control, and that is nearly as cost effective as todays sensor alone.
Until then, nearly all of this functionality can be contained at the transducer
device level. Advantech has been doing this for more than ten years and
continues research and development to reach these ultimate goals.
In part this comes through continual improvement of its ADAM transducer
technology. This allows multiple field signal types to be connected to a single
device, but also contributes to the size of the device. The larger package allows
for easier panel mounting, wiring and maintenance, and can contain more
processing and networking power. For instance, some years ago Advantech
added a three port switch to this small module allowing a daisy-chain topology.
This was so well received that this technology is now being added to its core

Modbus/TCP product. Obvious enhancements like switch pass through on

failure to allow connectivity for downstream modules or incorporation of SNMP
protocol

to

allow

remote

monitoring

of

communication

health.

Until

manufacturers are able to make the Moores Law leap to edge device with a cost
effective solution that brings transducing, network connectivity, and the first
level of control into that small package, the transducer level may be the arena
for that development for the foreseeable future.