3.5
Describe different network topologies
Basically, a topology is a shape; so, a network topology is the
shape of a network. There is, however, a big difference between a physical
network topology and a logical network topology. The physical network topology
represents how the network looks to your “naked eye.” In other words, the
physical network topology is the way the components are arranged. The logical
network topology represents how information flows through the network, which may
not be the same as how it looks to your “naked eye.” You should understand the
main network topologies and the difference between a physical network topology
and a logical one. You should be able to recognize them given a diagram,
schematic, or description. In the following paragraphs, I will discuss each
network topology in greater detail.
Multiprotocol Label Switching (MPLS) is a relatively new
mechanism used to create a logical network topology with no dependence on a
specific underlying protocol. Instead, it uses labels to make forwarding
decisions on packets and thereby offers you a tremendous amount of flexibility
in regard to network planning and prioritization of traffic. For example, you
could decide to give a specific customer's traffic a higher priority through
your network, just because that customer adds more to your “bottom line” than
other customers do. The specific configuration of MPLS is beyond the scope of
this book.
A point-to-point connection is not really so much of a network
topology as it is a piece of one. Today's networks generally consist of many
point-to-point and various other types of connections. Point
to point just means that the connection is active only for the sender and
the receiver and that there are no other computers or devices involved. In fact,
point-to-point connections are said to create communication that is not shared
because the only communication is between the sender and the receiver.
Point-to-point connections between network devices, such as switches or routers,
can provide for very efficient network communication. In fact, you may remember
from the earlier discussion that full-duplex communication requires
point-to-point connections. Figure 3.20 illustrates a point-to-point connection between two
routers.
Point-to-multipoint connections are created when an interface
is connected to two or more other interfaces. This is the general effect of a
hub on a network in which the data flows into one interface and can flow out of
all other interfaces. It can also be seen in router configurations, such as
Frame Relay switching (which I will discuss later in the chapter), in which the
point-to-multipoint connections are created using subinterfaces (virtual
interfaces). Point-to-multipoint Ethernet connections cannot use full-duplex
communications because the connections are shared and therefore require the use
of CSMA/CD to control the traffic. Point-to-multipoint connections in Frame
Relay switching might require the use of special protocols and configuration to
control data traffic. Figure 3.21 illustrates point-to-multipoint configurations.
A ring is a legacy topology that looks exactly like a star
topology to the naked eye. The real difference in a ring topology vs. a star
topology is the technology that is used. Computers in a ring topology generally
used IBM Token Ring technology. Other components can also be arranged in a ring
topology and use different technologies. The computers in a ring topology are
not generally arranged in a physical ring. In fact, just as with a star
topology, they can be next to each other or spread throughout a building. The
difference is that the central component that connects them contains the logical
ring that facilitates communication on the network using the ring technologies.
Figure 3.22 shows a ring topology. Please note that data flows
in a very different way, even though the physical topology would be
indistinguishable from that of a star topology to the naked eye.
A star topology is a group of computers connected to a central
location, such as a hub or a switch. This is the most common topology in use
today. The computers may be physically located next to each other or spread
throughout an entire building, but the flow of information from among computers
must go through the central location. In a star topology, each computer has its
own cable or connection to the hub. Since each computer has its own connection,
one computer's failing will not affect the other computers in the network.
However, if the hub or switch should fail, then all the computers on that hub or
switch will be affected. Figure 3.23 is an illustration of a star topology.
The full mesh topology is not often used for networks and is
almost never used for individual computers. In a full mesh topology, all the
components in the mesh have independent connections to all the other components
in the mesh. For example, if there were four computers connected with a full
mesh, then the number of connections could be determined by the following
formula:
[n(n-1)] / 2 = total number of connections
In this example, there would be a total of 12 connectors for 6
connections, and each computer would have to contain 3 network interface cards:
[4(4-1)] / 2 = 6
Any network with multiple or redundant connections to network
components can be considered a mesh topology, but because of the expense
involved in building this type of network, they are rarely created for
individual computers. A mesh, or even a full mesh, would most likely be found
connecting multiple networks in an organization. In fact, the Internet is the
best and biggest example of a partial mesh topology. Figure 3.24 shows a full mesh topology with four computers.
The bus topology was used in earlier networks but is not
commonly used today. In a bus topology, all computers are connected to each
other by a single cable. Coaxial cable with special connectors called BNC
connectors (as shown earlier in Figure 3.8) and T connectors were used. The T connectors provided an
independent connection for each computer on the bus. In addition, the bus worked
only if both ends of the cable had a special resistor, called a terminator, installed. Figure 3.25 shows a bus topology, and Figure 3.26 shows the T connector used to connect the computers
to the bus.
Very small business networks and home networks are often
peer-to-peer. This means that no dedicated server is involved at all. Each
computer acts as both a client and a server. Typically, directory shares or
folders are set up on each of the computers, and local accounts on the computers
are used to provide some minimal security. Generally, peer-to-peer networks
consist of no more than about 10 computers. A network of more than 10 computers
creates tremendous confusion because the users might have to know different
usernames and passwords to get to the share directories on each computer. Also,
what if the other nine computers wanted to use a share directory on the 10th
computer all at the same time? The 10th computer's resources would be so
overwhelmed with providing the share directory for the others that you might not
even be able to use it yourself!
In a client-server network, the problem of resource sharing is
addressed by using specific high-capacity and high-speed computers to share
resources to the client computers. Most of the resources that the clients use
are centralized to the very fast server computer. The server computer is
typically not used directly by a user. In the most sophisticated networks, these
servers are also domain controllers that authenticate a user's access onto the
network and control their access to specific resources.
Actually, most networks today are a combination of many
topologies. For example, a network will often use a star topology with a partial
mesh consisting of some point-to-point and some multipoint connections. This
type of hybrid design facilitates customization to the organization's
communication needs as well as redundant connections for load balancing and
fault tolerance. Figure 3.27 illustrates a hybrid network topology.
Know the difference between a physical topology
and a logical topology You should know that the physical topology of the
network is simply what it looks like or how the components are arranged. The
logical topology, on the other hand, represents the flow of information in the
network.
Know the main logical network topologies
You should be able to recognize the MPLS, point-to-point, ring, star, mesh, bus,
and hybrid topologies by a diagram, schematic, or description. You should be
able to recognize the difference between a peer-to-peer and a client-server
network
Basically, a topology is a shape; so, a network topology is the
shape of a network. There is, however, a big difference between a physical
network topology and a logical network topology. The physical network topology
represents how the network looks to your “naked eye.” In other words, the
physical network topology is the way the components are arranged. The logical
network topology represents how information flows through the network, which may
not be the same as how it looks to your “naked eye.” You should understand the
main network topologies and the difference between a physical network topology
and a logical one. You should be able to recognize them given a diagram,
schematic, or description. In the following paragraphs, I will discuss each
network topology in greater detail.MPLS
Multiprotocol Label Switching (MPLS) is a relatively new
mechanism used to create a logical network topology with no dependence on a
specific underlying protocol. Instead, it uses labels to make forwarding
decisions on packets and thereby offers you a tremendous amount of flexibility
in regard to network planning and prioritization of traffic. For example, you
could decide to give a specific customer's traffic a higher priority through
your network, just because that customer adds more to your “bottom line” than
other customers do. The specific configuration of MPLS is beyond the scope of
this book.Point to point
A point-to-point connection is not really so much of a network
topology as it is a piece of one. Today's networks generally consist of many
point-to-point and various other types of connections. Point
to point just means that the connection is active only for the sender and
the receiver and that there are no other computers or devices involved. In fact,
point-to-point connections are said to create communication that is not shared
because the only communication is between the sender and the receiver.
Point-to-point connections between network devices, such as switches or routers,
can provide for very efficient network communication. In fact, you may remember
from the earlier discussion that full-duplex communication requires
point-to-point connections. Figure 3.20 illustrates a point-to-point connection between two
routers. "Click To expand")
Point to Multipoint
Point-to-multipoint connections are created when an interface
is connected to two or more other interfaces. This is the general effect of a
hub on a network in which the data flows into one interface and can flow out of
all other interfaces. It can also be seen in router configurations, such as
Frame Relay switching (which I will discuss later in the chapter), in which the
point-to-multipoint connections are created using subinterfaces (virtual
interfaces). Point-to-multipoint Ethernet connections cannot use full-duplex
communications because the connections are shared and therefore require the use
of CSMA/CD to control the traffic. Point-to-multipoint connections in Frame
Relay switching might require the use of special protocols and configuration to
control data traffic. Figure 3.21 illustrates point-to-multipoint configurations. "Click To expand")
Ring
A ring is a legacy topology that looks exactly like a star
topology to the naked eye. The real difference in a ring topology vs. a star
topology is the technology that is used. Computers in a ring topology generally
used IBM Token Ring technology. Other components can also be arranged in a ring
topology and use different technologies. The computers in a ring topology are
not generally arranged in a physical ring. In fact, just as with a star
topology, they can be next to each other or spread throughout a building. The
difference is that the central component that connects them contains the logical
ring that facilitates communication on the network using the ring technologies.
Figure 3.22 shows a ring topology. Please note that data flows
in a very different way, even though the physical topology would be
indistinguishable from that of a star topology to the naked eye. "Click To expand")
Star
A star topology is a group of computers connected to a central
location, such as a hub or a switch. This is the most common topology in use
today. The computers may be physically located next to each other or spread
throughout an entire building, but the flow of information from among computers
must go through the central location. In a star topology, each computer has its
own cable or connection to the hub. Since each computer has its own connection,
one computer's failing will not affect the other computers in the network.
However, if the hub or switch should fail, then all the computers on that hub or
switch will be affected. Figure 3.23 is an illustration of a star topology. "Click To expand")
Mesh
The full mesh topology is not often used for networks and is
almost never used for individual computers. In a full mesh topology, all the
components in the mesh have independent connections to all the other components
in the mesh. For example, if there were four computers connected with a full
mesh, then the number of connections could be determined by the following
formula: [n(n-1)] / 2 = total number of connectionsIn this example, there would be a total of 12 connectors for 6
connections, and each computer would have to contain 3 network interface cards:
[4(4-1)] / 2 = 6Any network with multiple or redundant connections to network
components can be considered a mesh topology, but because of the expense
involved in building this type of network, they are rarely created for
individual computers. A mesh, or even a full mesh, would most likely be found
connecting multiple networks in an organization. In fact, the Internet is the
best and biggest example of a partial mesh topology. Figure 3.24 shows a full mesh topology with four computers. "Click To expand")
Bus
The bus topology was used in earlier networks but is not
commonly used today. In a bus topology, all computers are connected to each
other by a single cable. Coaxial cable with special connectors called BNC
connectors (as shown earlier in Figure 3.8) and T connectors were used. The T connectors provided an
independent connection for each computer on the bus. In addition, the bus worked
only if both ends of the cable had a special resistor, called a terminator, installed. Figure 3.25 shows a bus topology, and Figure 3.26 shows the T connector used to connect the computers
to the bus. "Click To expand")
Peer-to-peer
Very small business networks and home networks are often
peer-to-peer. This means that no dedicated server is involved at all. Each
computer acts as both a client and a server. Typically, directory shares or
folders are set up on each of the computers, and local accounts on the computers
are used to provide some minimal security. Generally, peer-to-peer networks
consist of no more than about 10 computers. A network of more than 10 computers
creates tremendous confusion because the users might have to know different
usernames and passwords to get to the share directories on each computer. Also,
what if the other nine computers wanted to use a share directory on the 10th
computer all at the same time? The 10th computer's resources would be so
overwhelmed with providing the share directory for the others that you might not
even be able to use it yourself!Client-server
In a client-server network, the problem of resource sharing is
addressed by using specific high-capacity and high-speed computers to share
resources to the client computers. Most of the resources that the clients use
are centralized to the very fast server computer. The server computer is
typically not used directly by a user. In the most sophisticated networks, these
servers are also domain controllers that authenticate a user's access onto the
network and control their access to specific resources.Hybrid
Actually, most networks today are a combination of many
topologies. For example, a network will often use a star topology with a partial
mesh consisting of some point-to-point and some multipoint connections. This
type of hybrid design facilitates customization to the organization's
communication needs as well as redundant connections for load balancing and
fault tolerance. Figure 3.27 illustrates a hybrid network topology. "Click To expand")
Exam Essentials
Know the difference between a physical topology
and a logical topology You should know that the physical topology of the
network is simply what it looks like or how the components are arranged. The
logical topology, on the other hand, represents the flow of information in the
network.Know the main logical network topologies
You should be able to recognize the MPLS, point-to-point, ring, star, mesh, bus,
and hybrid topologies by a diagram, schematic, or description. You should be
able to recognize the difference between a peer-to-peer and a client-server
network