You are on page 1of 12

LAN Topology

What will be the shape, or topology, that defines the LAN?


Topology is a term that refers the shape of the network and the layout of cabling from a bird's eye
view, much as a floor plan identifies the layout of offices and hallways in a building.
The topology diagram shown is made up of LAN workstations, also called nodes. The lines between
these nodes represent cables. Junction boxes are used to interconnect the cables. You will learn more
about the various types of hardware that support the connectivity function later in the course.
Theoretically, any kind of cable medium could be wired together to create an infinite variety of
network layouts. But, in fact, certain implementations of topology have become more popular while
other implementations have simply never caught on. Consider some of the most common types of LAN
topology.
The Star topology shown above is emerging as the most common network layout in use today. Each
workstation is connected point-to-point to a single central location that is commonly referred to as a
wiring closet.

LAN Bus Topology

The Bus topology is one of the oldest layout types and remains one of the simplest forms to design and
implement. Each seat in the network is connected in sequence along a single network cable using Tshaped network interface connectors, and terminating points are placed at each end.

LAN Mesh Topology


The Mesh topology is a variation of the bus, in which all devices are connected to one another in a
daisy-chain fashion, as opposed to connecting in sequence to a single network cable. Each node is
capable of transmitting, receiving, and routing data.

LAN Ring Topology


The Ring topology is another simple design that consists of a single cable that forms the main data
path in the shape of a ring. Each node is connected to a closed loop of cable. Signals travel in one
direction from one node to all other nodes around the loop.

LAN Tree Topology


The Tree topology is essentially a hybrid of the bus and star layouts. The basic topology is similar to
that of a bus, with nodes connected in sequence to a linear central cable. But tree networks may have
"branches" that contain multiple workstations that are connected point-to-point in a star-like pattern.
Signals from a transmitting node travel the length of the medium and are received by all other nodes.

Connection-oriented vs. Connectionless


Will the LAN be connection-oriented or connectionless?
Depending on the type of communication technology implemented, LANs can be either connectionoriented (such as an ATM network) or connectionless (broadcast-based, such as Ethernet).
A connection-oriented LAN requires that you establish contact with the receiving party before the
exchange of information can begin, in the same manner as you would for a telephone call. If you want
to call someone, you pick up the handset and dial the number. If, for whatever reason, there is no
answer at the other end, you do not bother to talk!
In the same way, a connection-oriented LAN will try to contact the node at the other end of an
intended conversation, and will only start sending information if the node at other end is present and
available. Once the other side has been contacted, it is not necessary to keep "redialing" the number,

because the connection to the other party has already been established. If the other end misses
something during the course of the exchange, they can request that the information be retransmitted.

Connection-oriented
Suppose you want to talk to two or more people. You could arrange a conference call by having the
other parties dial into a central place, such as an operator who would then join them together. If the
parties did not know where to call, they could not join. If the operator had a list of pre-approved
parties, then joining may have an element of security in place as well. The difference between a party
line and a conference call is the security element. You know when someone joins a conference call, but
you may not know who picks up on a party line.
Connection-oriented LANs also support two types of connections:

Point-to-point (private) connections refer to a pathway from one location to another (Node A
to Node B).

Point-to-multipoint (conference) connections refer to a pathway from one location to many


locations (Node A to Node B, C, and D).

Connectionless/Broadcast
Now, contrast connection-oriented communication with connectionless or broadcast-based
communication.
Broadcast-based communication is connectionless in that it does not require the transmitting party to
contact the receiving party before sending information. With a radio station, the news is broadcast
whether you are listening or not. If you were listening, but you missed something, you cannot ask the
announcer to go back and repeat what they just said. In the same way, a connectionless LAN simply
broadcasts information to all who may or may not be listening.
Some radios offer two-way communication, but you still have the eavesdropping problem associated
with any broadcast media. The advantage of a two-way broadcast is you can holler, "George, are you
out there?" and everyone tuned to the frequency will hear the request. If George is out there, he
answers and tells you where he is located. Radio stations also do this with contests, when they say "If
you are listening and are our seventh caller, you could win..."
Nodes in Ethernet LANs locate each other in the same way, using what is called an ARP (Address
Resolution Protocol) Broadcast. The originating node broadcasts an ARP request across the network
asking for the location of the target node. The target node responds to the broadcast by sending its
physical address back to the originating node so that information can be transmitted.
In all cases so far though, the broadcasts are one-way. Imagine how much more complicated it gets if
you could be transmitting as well as receiving on a broadcast medium. You would need something like a
CB or police scanner to check for messages for you on all frequencies. It would be easy to miss a call.

LAN Protocols
Which protocols will be used to define the way in which communication takes place across the LAN?
When sending a message from Node A to Node B on the network, there are several additional questions
that may come to mind:

How will the message arrive at the destination? Will it arrive piece by piece requiring an
acknowledgement for each or will it arrive all at once whether or not the receiving node is
ready?

How will the message be formatted (frames, packets, cells)?


How will Node A communicate with Node B?
Will the relationship be master/slave, client/server or peer-to-peer?
How will the various software programs communicate with each other?

Will the same software be required on both systems or can a sending system have one type of
software while the receiving software is of another type?

The answer to all of these questions is determined by protocols. Protocols are the rules that hardware
and software follow in order for devices to communicate with each other
Protocols consist of several functional parts. The answers to the questions above will become clearer as
you consider each part in detail, beginning with flow control.

Protocol Flow Control


How will the message arrive at the destination? Will it arrive piece by piece requiring an
acknowledgement for each or will it arrive all at once whether or not the receiving node is ready?
Protocols use different methods of flow control to monitor the pace at which data is transferred
between network nodes. There are three main types of flow control:

The request/reply method mandates that the receiving node acknowledge each data packet
before the next packet is sent.

The sliding window method allows multiple data packets to be transmitted between nodes
before acknowledgement. Both ends keep track of what has sent and what has been
acknowledged.
The unreliable method is one where there is no guarantee that the packet will arrive at its
destination. Data may be discarded at any time without notifying the sender or receiver.

Protocol Message Format


How will the message be formatted (frames, packets, cells)?
In communications, the format of information sent varies widely. In addition, the distinction between a
frame and packet is not always clear. While some network vendors attempt to differentiate the two,
the two terms are often used interchangeably. The term packet will be used throughout this course for
purposes of this training.
Both frames and packets are parts of a transmitted message that vary in size. The actual format of
may differ depending on the protocol or network specification. But in a LAN, environment packets
always contain both data and addressing information.
The graphic shown above represents the standard specification (IEEE 802.5) for an Ethernet packet,
which contains four fields that are common to most LAN protocol packet types:

The destination address indicates where the message is going.

The source address indicates where the message originated.


The data field contains the actual user information.
The frame check sequence assists in detecting errors within transmitted packets.

In addition, most packets contain a field (such as a preamble or delimiter) that is used to notify the
receiving station that the packet has arrived.

Protocol Message Format


Cells are fixed-sized (53 bytes) packets of data used to transfer information in an ATM network. The
fixed size of a cell allows for fast and efficient transfer of information across the network.
The cell format consists of two primary parts:

The header contains control information that is used to route and manage cells.

The payload contains the actual user information.

Packets and cells both rely on network protocols, as well as hardware and software, to determine how
they function in the LAN environment.

Network Architecture
How will Node A communicate with Node B? Are network communications based on a master/slave,
peer-to-peer, or client/server relationship?
There are three basic types of architectures that determine how one network node communicates with
another in a LAN:

Master-slave communications refer to a relationship where a single node ("master") initiates


and controls a session with one or more devices ("slaves"). Originally designed for mainframe
computer networks where the mainframe was the master and the "dumb" terminals were
slaves. The master/slave architecture is not commonly used in modern networks except in
isolated cases (For example, terminal emulation).

Peer-to-peer communications refer to a relationship where both nodes share the responsibility
of initiating, maintaining, and terminating a session. Due to performance limitations as
network load increases and a lack of administrative control, peer-to-peer architectures are
typically limited to small, single-platform, and low-traffic LAN environments.
Client/server communications refer to a relationship where dedicated servers support the
clients that are attached to them. Also known as two-tier and three-tier architectures,
client/server communications are commonly found in large, multi-platform, high performance
networks where security is an issue.

Protocol Stack
How will the various software programs communicate with each other? Will the same software be
required on both systems or can a sending system have one type of software while the receiving
software is of another type?
In order for two different computers to communicate (e.g. DOS and UNIX), the same protocol stack
must be in operation. The protocol stack defines the hierarchy, or ranked order, in which protocols are
applied to a communications network.
A widely used model for understanding network communications is the Open Systems Interconnection
(OSI) reference model. This model defines the protocols that allow network hardware and software to
work together. The OSI model determines how nodes contact each other, transmit data, and ensure
that transmissions are received, as well as, how they are connected. There are seven layers to the
model that define tasks that need to be completed (and protocols that are used to perform them).

Network Management
Once you have a LAN, how do you manage and troubleshoot the nodes?
Network management refers to how computer networks monitor and manage active functions,
including the following:
Fault management detects, displays, and maintains records of alarm conditions.

Configuration management handles additions, deletions, and changes to the network.

Accounting management tracks the measurable use and cost of network resources.

Performance management is designed to optimize the speed of the network.

Security management deals with such issues as logging on.

If you have a simple network with only three devices, then it might be just as easy to manage each
device individually. But if your network is larger and more complex, with dozens (maybe even
hundreds!) of devices that are logically grouped according to class, then a management tool might
make sense.
Network management tools often require huge workstations to run their software, requiring hundreds
of MB of RAM. That workstation is called the Network Management Station (NMS). The most common
type of NMS in use is Hewlett Packard's OpenView.
At the most basic level, the NMS determines whether or not objects on the LAN are "alive." If an object
"dies," you want to know about it right away. If a brand new object appears, you will want to know
about it right away as well.

LAN Applications
What types of applications does the LAN support?
Software applications are programs designed for the end user (spreadsheets, games, graphics). Sharing
applications is why most people install LANs.
Specialized applications include:

Vertical applications are specific to a particular industry (medical, government, education,


etc.).

Strategic applications are vital to business operations, including Oracle, SAP, and Lotus Notes.
Vestigial applications refer to critical business applications that pre-date the LAN, such as
MINX.

Representation of Networking Devices


Several terms have been used to describe networking devices, but there are also graphical
representations of these devices. These icons will appear in some of the graphics in the following

chapters. When looking at any network diagram, the symbols used are semi-standard. Please note that
the actual representation of network devices on diagrams may vary from vendor to vendor.
The following is a guide to the symbols that are used to represent the various kinds of LAN networking
devices within this course.

Hubs, which act as a common connection point for devices in a network, are represented by
small, unmarked rectangular boxes because they are often invisible to the network. While a
hub may be visible on the NMS, the hub is always invisible to data passing through it.

Switches, which filter and forward packets between LAN segments, are represented by a box
marked with an X.
Routers, which connect any number of LANs, are depicted by a square marked with a
unidirectional cross pattern.
File Servers, which provide a centralized file repository for LAN users, are depicted by files
overlaying a server device.
Database Servers, which support LAN database services, are depicted by a database bucket
overlaying a server device.
Network Management Stations (NMS), which supports LAN management services, is depicted
by an NMS terminal overlaying a server device.

Summary

This lesson provided an introduction to the basic terminology used to describe the physical
layout of a LAN, the protocol and communication standards that determine how information
is exchanged across the LAN, and various types of transmission media and networking
devices that are available to construct a LAN.

You might also like