Chapter 3 Switching Networks

 Ethernet (IEEE 802.3)

 Token Ring (IEEE 802.5)
 FDDI
 Frame Relay
 SMDS – Internetworking with SMDS
 Packet switching networks :
 TDS
 SDS
 Modular Switch design
 Distributed
 Shared
 Ethernet (IEEE 802.3)
Ethernet is the most popular LAN technology and specified by IEEE 802.3
standards published in 1985.
Ethernet used twisted pair wires or optical fiber cable.
10 BASE-T transmits at 10 Mbps and 100 BASE –T or fast Ethernet at
100 Mbps (1 gigabit or 1 billion bps)
Gigabit Ethernet , a speed of 1,000 Mbps with optical fiber for enterprise –
wide backbones.
Ethernet provides high throughput and low delays , for many applications.
Physical Layer :
The 10BASE –T wiring is shown in figure 3.1, it use the same category of UTPs
or UTP-3s.
In 10BASE –T networks each computer is equipped with a network Interface
card (NIC) and attached to a medium access unit(MAC) with two twisted pairs.
 Ethernet (IEEE 802.3)

Figure 3.1 Physical layout of a 10BASE-T network. This network uses UTP and can be wired
with telephone pairs already in place in the building. The transmission rate is 10 Mbps.
One can attach hubs together to build larger networks.
 Ethernet (IEEE 802.3)
• Similarly 100 BASE –T transceivers operates at 100 Mbps of UTP-3
• 1000 BASEx , 1000-LX is for transmission over single –or multimode fiber.
• 1000BASE-SX is for multimode fiber, 1000 BASE –CS is for copper shielded cables.
• 1000BASE –T will consider transmission over UTP.
Wireless Ethernet :
Wireless Ethernet , all the stations share a radio channel.
Physical layer standards specify the frequency spectrum(radio channel use) and
modulation scheme.
10 BASE 5
• 10 BASE 5 consists of segments.
• Each segment is a length of up to 500 m of coaxial cable with a diameter of 10mm and a
characteristic impedance of 50 Ohms.
• Segments are connected by repeaters that can be up to 1,000 m apart.
• No computers on the network can be more than 2,500 m apart.
• The transmission rate is 10 Mbps; the physical layer uses the Manchester encoding.
 Ethernet (IEEE 802.3)
• MAC – Medium Access Control:
• The MAC sublayer of Ethernet specifies the MAC addresses of network interfaces, the
frame format and the MAC protocol for sharing the cable.
• The frame format of Ethernet is shown in figure 3.2

DA- Destination Address

SA – Source Address
LEN – LENgth Indicator
PAD - Padding field
CRC –Cyclic redundancy Check
LLC- Logical Link control
DSAP – Destination service access points
Figure 3.2 Format of Ethernet packets. The various fields SSAP – Source service acces points
that make up the frame are explained in the abbreviation .

• Ethernet MAC protocol is CSMA / CD, carrier Sense Multiple Access

with Collision Detection.
 Ethernet (IEEE 802.3)

Figure 3.4 Sequence of events on an Ethernet network.

The transmitters waste five propagation times per
Figure 3.3 The maximum time until a node detects a collision is successful transmission , on the average.
twice the propagation time of a signal between the nodes that are
farthest apart.

 PROP is the propagation time of a signal from one end of the cable to the other.
 Accordingly , the fraction of time that the nodes use the transmission channel to transmits packets
successfully approximately equal to 1 / (1 + 5a) where “a” denotes the ratio of a propagation time (PROP) to
a packet transmission time(TRANS) .
 This fraction of useful time when many nodes want to transmit is called ‘efficiency’ of the
MAC.
• LLC – Logic Link Control : Ethernet (IEEE 802.3)
IEEE 802.2 LLC standards is used for 802.3,802.5 and other networks.
LLC sublayer provides connection-oriented or connectionless (acknowledged or not) services.
LLC also multiplex different transmissions that are differentiated by the service access point
(SAP) field as shown in figure 3.2
LLC implements the transparent routing of packets between Ethernets attached together
with bridges as shown in figure 3.5. (LAN Interconnection with Bridge)
LLC uses CRC field to detect errors, LLC with Go-back N protocol to arrange for the
transmitter to retransmit packets that do not arrive error free.

Figure 3.5 Ethernet attached by bridges. Transport

routing enables a computer to transmit a packet to
another computer on one these Ethernets as if it were
on the same Ethernet.
 Token Ring (IEEE 802.5)
• The IEEE 802.5 standards specify layers 1 and 2 of a family of token ring
networks.
• It transmit at 4 Mbps or 16 Mbps and developed for 100-Mbps transmission
speed.
• Token ring advantage is unlike Ethernet, each node is guaranteed to be
allowed to transmit before a specific time.
• Token ring are more efficient than Ethernet under high load.
• Token Ring - Physical layer:
• Network are connected into a ring by point-to-point links see in figure 3.6
• TR – network interface - Two possible configurations:
• (a) Repeater – In this configuration interface repeats the incoming signal on
the outgoing link with a delay of a few bit transmission times.
• (b) Open – In open configuration , the interface transmits on the outgoing link
and listen on the incoming link.
 Token Ring (IEEE 802.5)

Figure 3.6 Layout of a token ring network. The Figure 3.7 steps in the transmission of a packet
computers are attached by unidirectional point-to- when the computers use the release after
point links around a ring. transmission token-passing protocol.

Token ring – MAC :

 The frame format is similar to that of Ethernet packets (figure 3.2) , except it uses an ending delimiter instead
of a length indication.
 The transmission proceed in one direction along the ring.
 Figure 3.7 shows the sequence of events when a node wants to transmit a packet with following actions.
a. Release after transmission b. Release after reception c. token holding time (THT)
 Token Ring (IEEE 802.5)
• The timing diagram of Token Ring as shown in figure 3.8

Figure 3.8 Timing diagram for the release

after reception token-passing protocol when
all the computers have packets to transmit.

 The figure shows the efficiency is approximately equal to 1/ (1+a) , where a = PROP / E( Tn).
 In this expression, E(Tn) is the average duration of a node transmission.
 Token ring efficiency is more than 90%
Note : TR-LLC of IEEE 802.5 networks is the same as in IEEE 802.3 networks
 FDDI - Fiber Distributed Data Interface
• FDDI is an ANSI (American National Standards Institute) standard for a
100 Mbps in 1987.
• Recently preferred technology for connecting LANs.
• FDDI connects up to 500 nodes with optical fibers , in dual topology.
• Distance between two nodes cannot exceed 2 Km when multimode fibers
and LEDs are used.
• Longer separation is possible with single mode fibers and laser diodes.
• Maximum length of the fibers 200km, with this length FDDI networks are
used to interconnect computers within a campus.
• the figure 3.9 shows the FDDI networks , connect workstations to file servers
and printers , or work station together , or terminals and terminal emulators
workstations.
 FDDI - Fiber Distributed Data Interface

Figure 3.9 An FDDI dual ring network can support 500 stations with a total distance of 200Km and up tp 2km between
adjacent stations .Stations are connected by 100-Mbps optical fiber , and a timed –token MAC protocol is used
 FDDI - Fiber Distributed Data Interface

Figure 3.11 when a fault is detected, the rings are

configured to isolate the faulted station.

Figure 3.10 The FDDI standards specify the MAC

sublayer and the physical layer of the protocol stock
 FDDI - Fiber Distributed Data Interface
• Figure 3.10 , specify the FDDI standards the MAC sublayer and physical layer.
• Physical layer itself is divided into two sublayers, with station management (SMT)
protocols.
• PMD sublayer specifies the fiber to be used with optical source and detectors.
• PHY (physical) sublayer specifies the station must use 4B / 5B encoding..
• The SMT must detect errors and isolate a fault on the ring, such as failure of a
station or link on the ring.
• Figure 3.11 illustrates how the dual ring is reconfigured as a single ring after the fault
has been isolated.
• In addition, the SMT monitors the performance of the network.
• MAC of FDDI specifies the frames have a maximum length of 4,500 bytes.
• Frame structure is illustrated in figure 3.10.
• MAC uses a time-token protocol.
• This protocol is similar to the token –passing mechanism of IEEE 802.5 , except for
the timing feature , it will be explain next.,
 FDDI - Fiber Distributed Data Interface
Figure 3.12 The timing token guarantees that each station will get a
chance to transmit in less than TTRT. TTRT is the target token rotation
time

Figure 3.12 helps to explain the MAC protocol when the station transmit only asynchronous traffic.

Analysis of MAC protocol in FDDI :

 Assume the stations are initially idle, i.e. they have no packet to transmit.
 A token , which is a packet with a specific bit pattern, travels around the ring.
 Each station has two timers : TRT or token rotation time timer, which counts up, and THT or
token holding time timer, which counts down.
 When a station has a packet to transmit, it waits until it gets the token.
 When the station gets the token, it does the following.
1. Grabs the token.
2. Sets THT = TTRT – TRT ( TTRT or target token rotation time is set by the network manager)
3. Resets TRT = 0
4. Transmits packets until THT = 0 or there is no packet left.
5. Release the token.
 Frame Relay (FR)
• Frame Relay is a wide area network with the following features:
• 1.FR operates at a higher speed (1.544 Mbps and recently 44.376 Mbps). This means that it
can easily be used instead of a mesh of T-I or T-3 lines.
• 2. FR operates in just the physical and data link layers. This means it can easily be used as a
backbone network to provide services to protocols that already have a network layer
protocol, such as the Internet.
• 3. FR allows bursty data.
• 4. FR allows a frame size of 9000 bytes, which can accommodate all local area network
frame sizes.
• 5. FR is less expensive than other traditional WANs.
• 6. FR has error detection at the data link layer only.
• 7. There is no flow control or error control. There is not even a retransmission policy if a
frame is damaged; it is silently dropped.
• 8.FR was designed in this way to provide fast transmission capability for more reliable
media and for those protocols that have flow and error control at the higher layers.
 Frame Relay (FR)
• FR Architecture
• FR provides permanent virtual circuits and switched virtual circuits.
• Figure 3.13 shows an example of a Frame Relay network connected to the Internet.
• The routers are used to connect LANs and WANs in the Internet.
• In the figure, the Frame Relay WAN is used as one link in the global Internet.
• Virtual Circuits Frame Relay is a virtual circuit network.
• A virtual circuit in Frame Relay is identified by a number called a data link connection
identifier (DLCI).
• VCIs in Frame Relay are called DLCIs.

Figure 3.13 Frame Relay Network

 Frame Relay (FR)
• Frame Relay Layers :
• Figure 3.14 shows the Frame Relay layers. Frame Relay has only physical and data link
layers.
• Frame Relay operates only at the physical and data link layers
• Physical Layer :
• No specific protocol is defined for the physical layer in Frame Relay. Instead, it is left to
the implementer to use whatever is available.
• Frame Relay supports any of the protocols recognized by ANSI.
• Data Link Layer :
• At the data link layer, Frame Relay uses a simple protocol that does not support flow or
error control. It only has an error detection mechanism.

Figure 3.14 Frame Relay layers

•
 Frame Relay (FR)
• Frame Relay frame
• Figure 3.15 shows the format of a Frame Relay frame. The address field
defines the DLCI as well as some bits used to control congestion.
• The descriptions of the fields are as follows:

Figure 3.15 Frame Relay frame

 Frame Relay (FR)
• Address (DLCI) field. The first 6 bits of the first byte makes up the first part of the DLCI.
The second part of the DLCI uses the first 4 bits of the second byte.
• These bits are part of the lO-bit data link connection identifier defined by the standard.
• Command/response (CIR). The command/response (C/R) bit is provided to allow upper
layers to identify a frame as either a command or a response.
• It is not used by the Frame Relay protocol.
• Extended address (EA). The extended address (EA) bit indicates whether the current byte
is the final byte of the address.
• An EA of 0 means that another address byte is to follow (extended addressing is
discussed later).
• An EA of 1 means that the current byte is the final one.
• Forward explicit congestion notification (FECN). The forward explicit congestion
notification (FECN) bit can be set by any switch to indicate that traffic is congested.
• This bit informs the destination that congestion has occurred.
• In this way, the destination knows that it should expect delay or a loss of packets.
 Frame Relay (FR)
• Backward explicit congestion notification (BECN). The backward explicit
congestion notification (BECN) bit is set (in frames that travel in the other
direction) to indicate a congestion problem in the network.
• This bit informs the sender that congestion has occurred. In this way, the
source knows it needs to slow down to prevent the loss of packets.

• Discard eligibility (DE). The discard eligibility (DE) bit indicates the priority
level of the frame. In emergency situations, switches may have to discard
frames to relieve bottlenecks and keep the network from collapsing due to
overload.
• When set (DE 1), this bit tells the network to discard this frame if there is
congestion. This bit can be set either by the sender of the frames (user) or
by any switch in the network.

• Frame Relay does not provide flow or error control; they must be provided
by the upper-layer protocols.
 Frame Relay (FR)
• Extended Address :
• To increase the range of DLCIs, the Frame Relay address has been extended
from the original 2-byte address to 3- or 4-byte addresses.
• Figure3.16 shows the different addresses.
• Note that the EA field defines the number of bytes; it is 1 in the last byte of
the address, and it is ain the other bytes.
• Note that in the 3- and 4-byte formats, the bit before the last bit is set to O.

Figure 3.16 Three address formats

 Frame Relay (FR)
• FRADs
• To handle frames arriving from other protocols, Frame Relay uses a device called a Frame Relay
assembler/disassembler (FRAD).
• A FRAD assembles and disassembles frames coming from other protocols to allow them to be carried by
Frame Relay frames.
• A FRAD can be implemented as a separate device or as part of a switch.
• Figure 3.17 shows two FRADs connected to a Frame Relay network.
VOFR
• Frame Relay networks offer an option called Voice Over Frame Relay (VOFR) that sends voice through the
network.
• Voice is digitized using PCM and then compressed. The result is sent as data frames over the network.
• This feature allows the inexpensive sending of voice over long distances.
• However, note that the quality of voice is not as good as voice over a circuit-switched network such as the
telephone network.
• Also, the varying delay mentioned earlier sometimes corrupts real-time voice.

Figure 3.17 FRAD

 SMDS
• Switched Multimegabit Data Service (SMDS) is a public switched connectionless data
transport service , defined by Bellcore.(BOC-Bellcore operating Companies)
• SMDS offered at follows :
• DS-1(T1) access speed (1.54 Mbps line rate corresponding to 1.7 Mbps line rate )
• DS-3(T-3) access speed (45 Mbps line rate or 34 Mbps data rate)
• SMDS later be offered higher rates over a SONET Network.
• Figure 3.18 is a SMDS network via dedicated access lines.

Figure 3.18 The SMDS Network and interfaces

 SMDS
• SMDS protocol roughly corresponds to the first three OSI layers.
• It is divided into three levels.
• Level 1 (L1) : provides the physical interface to the digital network.
• Level 2 (L2) : defines a cell structure similar to ATM cells and performs error
detection.
• Level 3 (L3) : handles addressing and routing.
• Figure 3.19 Shows the formats of protocol data units (PDU) at levels 2 and 3.
• User data up to 9,188 bytes, is an encapsulated in an L3_PDU.
• The L3_PDU overhead contains the full source and destination addresses, the
L3_PDU length and , optionally , a CRC for detecting L3_PDU errors.
• Each address is specified by 15 BCD( Binary coded decimal) digits.
 SMDS

Figure 3.19 The SMDS protocol stack and frame structure.

 Internetworking with SMDS
• SMDS and Frame relay are used to interconnect LANs.
• Conclude the functions that users connected to different FDDI LANs can transparently
interconnect over an SMDS network.(The functions are similar if Frame relay is used).
• Figure 3. 20 shows user 1 on one FDDI ring who wishes to send a packet to user 2 on
another ring..

Figure 3.20 FDDI networks can be interconnected using SMDS

 Internetworking with SMDS
Analysis of Operation for Internetworking with SMDS :
• Assume first that user 1 knows the MAC address of user 2.
• In that case user 1 places a frame on FDDI 1 with destination address of user 2.
• Action 1 : Bridge b1 on that ring must:
• 1. copy that frame;
• 2.”convert” it into an L3_PDU , address it to station b2(using its SMDS address),
and submit it to the SMDS network.
• Note that bridge b1 is a station on FDDI 1 and on the SMDS network.
• In order to “submit” the L3_PDU, it must go through the SMDS protocol stack,
fragmenting the L3_PDU in to L2_PDUs, and then organizing them for
transmission into L1_PDU as in Figure 3.19.
• When bridge b2 receives these L1_PDUs, it assembles them in to the L3_PDU.
 Internetworking with SMDS
• Subsequently,
1. Bridge b2 must “convert” the L3_PDU back in to the original FDDI frame ,
place it on FDDI 2 , and then remove it from that ring when it returns
(note that the source address on the frame is that of user 1 and not b2);
2. User 2 copies this FDDI frame.

Action 2 : Similarly, b2 is on FDDI 2 and on the SMDS network.

 Internetworking with SMDS
• The most common way to do this operation to encapsulate the entire
FDDI frame as a payload in an L3_PDU , which is shown in the top panel
of Figure 3.21.

Figure 3.21 internetworking with SMDS

 Packet Switching Networks
• Switch / Switching in Networks :
• Networks use switching to achieve connectivity among users while sharing
communication links.
• Switches networks can be establish higher-capacity communication paths
among users with fewer links at lower per user cost.
• In this way networks can advantage of the economics of scale in
communication links.
• Switches a device (black box) with several input and output ports ,
terminate the incoming and outgoing links.
• An incoming links carries multiplexed bit streams of several users.
• Switches guides bit streams to the output port.
 Switch / Switching in Networks :
• Switch Performance Measures :
• Switch as a “ black box” .
• It is a device M input links and N output links as in the top of the figure 3.22

Figure 3.22 Switch as a “black box”(top) ;

a cross bar switch (middle) ; and a multipoint
switch (bottom)
 Switch / Switching in Networks :
• The bits or packets arrive on one link leave to another designated link.
• Performance measures can be to used to compare switches : connectivity, delay, set up
time, throughput, and complexity.
• Connectivity :- it is measured by set of pairs of input and output links simultaneously
connected through the switch.
• It process the incoming bit streams in order to route to the proper output ports .
• Delay is another measure of performance.
• Packet switches introduces queuing delay.
• Throughput of a switch is measured in terms of number of ports and the speed of the
individual input links.
• Measures to estimate complexity of a switch include the number of cross points , buffer
size , and speed of bit streams inside the switch.
• Switch design needs expensive electronic components.
 Switch / Switching in Networks :
• Crossbar switch , every pair of input and output links is connected by a cross point as in the
middle of figure 3.22.
• Input –output pairs (1,2),(2,1) and (M,N) are connected ( A dark point means it is closed).
• Proposition 1: ( N X N point-to-point switch) :
If a switch is a collection of cross points , then
number of cross points ≥ 𝑙𝑜𝑔2 (number of different configurations)
• Proposition 1 implies the number of cross points of a switch is at least
𝑙𝑜𝑔2 (N!) ≈ N 𝑙𝑜𝑔2 𝑁 .
Using Stirling's approximations :
𝑁 𝑁
N!≈ ( ) , Where e = 2.718
𝑒
• Therefore, not possible to build an N X N point-to-point switch with fewer than N 𝑙𝑜𝑔2 𝑁
cross points.
• The degree of blocking in switches with fewer than number of cross points.
 Switch / Switching in Networks :
• In the case of N X N multipoint switch ( see bottom of figure 3.22)

• The input link to any set of output links.

• The only restriction on the configurations is that different input links connected to
different output links .

• (This type switch is multicast capability , useful for video conferencing and video
distribution).

• This switch has 𝑵𝑵 possible configurations.

• In summary , an NX N point –to point or multipoint switch requires at least

N 𝑙𝑜𝑔2 𝑁 cross points
 Time-and –Space Division Switching – TDS & SDS
• The telephone network uses time-division switches (TDS) and figure 3.23
illustrates the operation of TDS.

Figure 3.23 A time-division switch consists of a

time-division multiplexer (MPX) , a time slot
interchanger , and a time-division
demultiplexer(DMX).
 Time-and –Space Division Switching – TDS & SDS
• Figure 3.23 top part shows N input signals arrive on N different links.
• Periodic bit streams go out on different output links.
• TDS has three parts : MPX , TSI and DMX.
• The bottom part of the figure 3.23 shows the delays of the signals in a TDS.
• Observe that the bit rate of the TSI is rate as the throughput of the switch., i.e.
N X b bps , where ‘b’ is the bit rate of each link.
• Therefore TDS switch is limited by the speed of the electronic components.
• To achieve greater throughput than possible by TDS alone, one can combine
it with space division.
• Comparative operation effect of TDS and SDS:
• TDS separates signals in TIME , a SDS separates the signals in SPACE.
Time-and –Space Division Switching – TDS & SDS
• Combine TDS and SDS as shown in figure 3.24.
• The boxes labelled ‘T’ are TDS.

Figure 3.24 SDS and TDS switching can be combined to produce switches.
 Modular Switch Designs

• Clos Network - Theorem and Analysis - Tutorial in Board

 Distributed Buffer Switch - DBS

Figure : Four variants of distributed buffer switches. Each 2 x 2 module contains an input or output buffer.
Distributed Buffer Switch - DBS

Figure :In self routing, the tag is decoded bit by bit by each module
to determine the route through the distributed buffer switch.
 Distributed Buffer Switch - DBS

Figure : Queuing in a banyan network :

two packets destined for different ports
Figure : Self routing in an omega network. may contend for transfer along the same
If the output port is ABC, the tag is CBA internal link causing a queuing delay