2 Link Layer Technologies

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 1 of 18

2.1 Introduction ..................................................................................3

2.2 Ethernet (1/4) ................................................................................4
2.2 Ethernet (2/4) ................................................................................5
2.2 Ethernet (3/4) ................................................................................6
2.2 Ethernet (3/4) ................................................................................7
2.3 Token Ring (1/3) ...........................................................................8
2.3 Token Ring (2/3) ...........................................................................9
2.3 Token Ring (3/3) ......................................................................... 10
2.4 Fiber Distributed Data Interface (FDDI) (1/3) ............................ 11
2.4 Fiber Distributed Data Interface (FDDI) (2/3) ............................ 12
2.4 Fiber Distributed Data Interface (FDDI) (3/3) ............................ 13
2.5 Frame Relay (1/2)........................................................................ 14
2.5 Frame Relay (2/2)........................................................................ 15
2.6 ATM (1/3) ..................................................................................... 16
2.6 ATM (2/3) ..................................................................................... 17
2.6 ATM (3/3) ..................................................................................... 18

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 2 of 18

2.1 Introduction

Different Link layer technologies are used to enable communication between computers on a
network. One difference between these technologies is the rules that apply when data is sent
to or received from the network cable.
This process, also called the access method, regulates and controls the data flow within the
network.
Before being sent over the transmission medium, a header and trailer are added to each
datagram. These fulfil several functions:
To divide the individual packets, the beginning and end of each packet are marked,
and the payload information is separated from the header and the trailer.
Furthermore, the higher protocol that is to receive the data is identified.
The source and destination of the data are identified with unique addresses.
Finally, a bit-stream level integrity check in form of a checksum guarantees the
identification of bit errors within the entire packet received by the hardware.
As was already mentioned in the previous chapter, Ethernet, Token Ring, FDDI, Frame
Relay and ATM are some of the numerous network technologies. We will examine these in
greater detail in the following lessons. Let's start with Ethernet.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 3 of 18

2.2 Ethernet (1/4)

Ethernet is a passive technology. Thus a network failure is only possible if the cable is
disconnected. Ethernet uses the Carrier Sense Multiple Access with Collision Detection
access method, or CSMA/ CD. CSMA/ CD is a system of rules which specifies how network
devices must react if two of them are trying to transmit data over the network at the same
time.
If several computers try to transmit data over the network at the same time, a collision
occurs. Each computer checks the transmission medium for network traffic. Data is sent only
after recognizing that the cable is available, that means that there is no network traffic. After
the data has been sent off, other computers that want to transmit data have to wait until the
data of the previous computer has reached its destination, and the cable is available again. If
a collision is detected, the network device waits for a random time interval, and then tries to
send the data again.
If another collision is detected after this interval, it waits for twice the time before resending
the data. As a standard, Ethernet operates at a speed of 10 Mbit/s. A faster version, called
Fast Ethernet, supports 100 Mbit/s, and Gigabit Ethernet even up to 1000 Mbit/s. Now, let's
have a look at the Ethernet frame format.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 4 of 18

2.2 Ethernet (2/4)

The Ethernet header includes the following fields:

The Preamble field is 8 bytes long and comprises 7 bytes of alternating 1- and 0-bits
for receiver synchronization, as well as a 1 byte sequence ending on 1 instead of 0.
This sequence marks the beginning of the data packet.
The Destination Address field is 6 bytes long. It can either be a unicast, a multicast or
a broadcast address. The unicast addresses, that is the point-to-point addresses, are
the MAC addresses we have already mentioned. The destination address is always
given in hexadecimal numbers.
The Source Address field is also 6 bytes long and contains the hexadecimal unicast
address of the source computer.
The Ethernet Type field is 2 bytes long and specifies the protocol of the higher layer
within the Ethernet frame.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 5 of 18

2.2 Ethernet (3/4)

The payload data is the data that has been handed down from the higher-level protocol.
Ethernet transmits payload data of a maximum size of 1500 bytes. For a reliable support of
collision detection, the minimum size of Ethernet packets is 46 bytes. The Frame Checksum
field is 4 bytes long and is used to verify the integrity of the data at bit stream level.
In contrast to other network technologies, Ethernet has no way of explicitly marking the end
of a data packet. Therefore it uses gaps to separate the data packets. Let's have a closer
look at the frame checksum.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 6 of 18

2.2 Ethernet (3/4)

The frame checksum is also called FCS - Frame Check Sequence - or CRC, Cyclical
Redundancy Check.
The source computer calculates the value and enters the result into the Frame Checksum
field. When the destination computer receives the packet, it applies the same algorithm and
compares its own result with the value inserted by the source computer. If the values match,
the data packet is accepted and processed. If the values don't match, the packet is deleted.
To calculate the CRC, a 33 bit long prime number is divided by the sum of all bits in the data
packet, leaving out the preamble and the CRC field. The result of this division is a quotient
and a remainder. The remainder is always a 32 bit value, and is inserted into the CRC field.
This way, bit errors can be detected with a probability of 100%.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 7 of 18

2.3 Token Ring (1/3)

Token Ring is another technology of the Link layer. Token Ring networks are implemented
as a ring topology. This ring is also called a logical ring. The physical topology of a Token
Ring network is star-shaped, with all computers connected physically to each other.
The access method used in a Token Ring network is known as token passing.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 8 of 18

2.3 Token Ring (2/3)

A token is a specific bit sequence which is passed around the network. Only the computer
that holds the token can transmit data.
As soon as the first computer on the Token Ring network goes online, the network generates
a token. This token is passed around to every computer on the ring, until one of the
computers takes control of it.
As soon as a computer has captured the token, it sends out a data frame to the network.
This frame is passed around the ring until it reaches the computer whose address matches
with the destination address of the frame. The frame is then copied by the destination
computer and marked to indicate that the data has reached its destination. Then the frame is
passed around the ring again, until it gets back to the source computer, where the
transmission is confirmed as successful before the frame is deleted. Afterwards, the source
computer sends out a new token to the ring, so the other computers can transmit their data
as well.
Token Ring networks typically operate at 4 or at 16 Mbit/s.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 9 of 18

2.3 Token Ring (3/3)

Taking a closer look at the Token Ring frame format, we can clearly relate the individual
bytes to different functions.
The start and end delimiters are each 1 byte long, and mark the beginning and end of a data
packet. The access control field is 8 bits long and contains information about priority and
token reservation. It also indicates if a token has already circulated the ring, and
distinguishes a frame from a token.
The bytes of the frame control indicate MAC control functions such as ring purge, beacon
and claim token.
The destination and source addresses are again the respective MAC addresses as
hexadecimal numbers.
In contrast to Ethernet, there is no minimum packet size for payload data. The maximum size
depends on a series of factors, such as the bit rate and the token hold timer. The frame
checksum is calculated like in Ethernet, that is on the basis of a 33 bit prime number.
Finally, the frame status field shows if the destination address has been recognized and
whether the packet has been successfully copied to the buffer of the network adapter.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 10 of 18

2.4 Fiber Distributed Data Interface (FDDI) (1/3)

An FDDI - Fiber Distributed Data Interface network - provides high-speed connections for
various network types. FDDI was designed for computers which require higher speeds than
the 10 Mbit/s provided by Standard Ethernet or the 16 Mbit/s provided by Token Ring
technology.
An FDDI network consists of two similar data streams flowing in opposite directions on two
separate rings. One ring is called the primary ring, the other one is known as the secondary
ring. If a problem arises in the primary ring, such as a ring error or a cable break, the data is
ent to the secondary ring, which continues the transmission.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 11 of 18

2.4 Fiber Distributed Data Interface (FDDI) (2/3)

Like a Token Ring, an FDDI network uses the token passing access method. However, a
computer on an FDDI network can transmit as many packets as it can generate within a
certain time determined by the Token Hold Timer. After expiration of the Token Hold Timer
the token must be released.
Since a computer does not release the token before the THT has expired, several packets
can circulate the ring simultaneously. This type of token passing is more efficient than in
standard Token Rings, which only allow the circulation of one packet at a time.
The optic fibre cables allow transmission speeds between 155 and 622 Mbit/s.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 12 of 18

2.4 Fiber Distributed Data Interface (FDDI) (3/3)

With Token Ring and FDDI using the same access method, it will come as no surprise that
the frame formats are similar as well. Essentially, there are three differences:
First, FDDI frames have no access control, but a preamble for receiver synchronization.
FDDI allows a maximum payload size of 4,500 bytes.
The third difference lies in the frame status field: It is 2 bytes long and contains indicators for
address identification, packet copying and beacon.
In contrast to the Token Ring technology, an error indicator notifies whether a packet must
be retransmitted.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 13 of 18

2.5 Frame Relay (1/2)

Frame Relay is a packet switching technology which transmits frames of variable length
across networks. These frames contain additional addressing and error handling information.
Communication is effected over a network which offers a permanent virtual data path
between the computers.
Typically, digital connections or optical fiber links are used, which provide quick access to the
data.
Packet switching is a method where a large amount of data is subdivided into smaller
packets. These packets are sent over a switch, which picks out the best route available at
the moment for the packets. Although the packets may reach their destination via different
paths, the receiving computer can reassemble them into the original chunk of data.
But there is also the chance to install a PVC (Permanent Virtual Circuit), which sends all the
packets over the same path.
This enables a faster data transmission, since, by chosing an appropriate route, the packets
do not have to be fragmented. Frame Relay is a point-to-point access technology, whose
transmission speeds depend on the trunk circuit.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 14 of 18

2.5 Frame Relay (2/2)

The header and trailer of a Frame Relay frame each contain a flag delimiting the beginning
and end of a frame.
The address field is 2 bytes long and contains data for the virtual connection as well as
indicators for congestion notification.
Furthermore, a control field guarantees the packet order and acknowledgement, and thus
ensures transmission reliability at bit level. Finally, the Network Layer Protocol Identifier field,
or NLPID, indicates which Internet layer protocol was used to encapsulate the payload data.
The Frame Check Sequence has already been discussed in the previous chapters.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 15 of 18

2.6 ATM (1/3)

The last Link layer technology we will discuss in this chapter is ATM - Asynchronous Transfer
Mode. ATM is also a packet switching technology, just like Frame Relay. But, in contrast to
Frame Relay, it transmits packets with a fixed length. Packets whose lengths are fixed, also
known as "cells", are data packets that only contain basic routing information, enabling
special devices - the switches - to forward the packets quickly to their destination.
Communication is effected over a point-to-point system, which offers a Permanent Virtual
Circuit between the computers.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 16 of 18

2.6 ATM (2/3)

Since ATM transmits packets of a fixed length and with little routing information using the
point-to-point method, high transmission speeds are achieved. Therefore, ATM also supports
the transmission of speech, realtime video, graphics, CD quality audio and bit rates in the
megabit range.
ATM networks typically support bit rates between 155 and 622 Mbit/s. Let's consider an ATM
cell in somewhat greater detail.

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 17 of 18

2.6 ATM (3/3)

As we have seen, ATM offers connection-oriented, but unreliable transmission services. The
responsibility for transmission reliability is taken over by a higher protocol, such as TCP.
ATM takes a random sized data packet, e.g. an IP datagram, and splits it up into 48 byte
fragments. Then each of these fragments has a 5 byte ATM header added to it. This total of
53 bytes is known as the ATM cell.
The 5 header bytes encompass a Virtual Path Identifier (VPI) supporting up to 4,096 virtual
paths per connection.
The 16 bits of the Virtual Channel Identifier can assign up to 65,536 separate virtual
channels to each virtual path.
The three following bits comprise the Payload Type Identifier. They contain information on
the cell type, congestion notification and segmentation. Another bit indicates whether the cell
is high or low priority.
Finally, the Header Error Check indicates possible bit errors within the first 4 bytes of the
ATM header. It only verifies the integrity of the header, not of the payload. Payload
verification is taken over by a higher layer protoco

Version 2.1

T.O.P. BusinessInteractive GmbH

Page 18 of 18