Computer Science

Networks
Section 1
Introduction
• Networks and topologies
• LANs and WANs
• Performance
• Client-server and peer-to-peer, thick and thin clients
• Network hardware and Ethernet
• The internet and WWW (and DNS)
• Cloud Computing
• Bus, star, mesh and hybrid topologies
• Wired and wireless networks, protocols, and layers
• Wired and wireless
• Encryption
• Addressing (IP and MAC)
• CIDR and Subnets
• Bitstreaming
What is a network?
• Two or more computers…
or other devices
• … connected together…
• … via a shared 'medium'…
which could be wired or wireless
• … for the purpose of communication
Benefits
• Allow the sharing of files/data. ​
• Allow communication between devices. ​
• Allow the sharing of resources e.g. hardware / software
(applications). ​
• Allow central management / backup, security, etc.
 Computer Computer Computer Computer Computer Computer Computer Computer

Shared Medium

Rules are required to govern access to the shared medium:

How do we ensure that all of the computers have a 'fair' share of access?
How do we resolve conflicts?

Rules are also required to govern the communication between computers:

How do we find the 'address' of the computer we wish to talk to?
How do we hold a 'conversation' with this computer?

A set of rules on a specific subject is called a protocol. TCP, HTTP, IP, FTP, SMTP are all Internet protocols.

Standards are required to ensure that systems can work together, even if they are built by different teams or companies.
A protocol is a particular kind of standard, others standards relate to hardware (e.g., the USB connector) and software
(e.g., the language C#).
These standards and protocols apply across networks of networks….

Computer Computer Computer Computer Computer Computer

Router Router

Shared Medium Shared Medium Shared

These standards and protocols apply across networks of networks….
… extending to billions of devices on the internet today.

Computer Computer Computer Computer Computer Computer

Router Router

Shared Medium Shared Medium Sha

Shared Medium Share

LAN/WAN
• Local Area Network – all of the connected devices are within a single,
relatively small geographical area. Often associated with an
organisation such as a small business, a school, or a home. Usually
owned and administered by a single organisation.
• Wide Area Network – a network of networks covering a large
geographical area, possibly across countries and continents. Owned
and administered by several cooperating organisations (some of
which may private/commercial, others may be public/government-
owned).
 LANs and WANs C C C S

LAN – Birmingham Office

R
UK WAN

R R
R

C C C R

R C C C S
LAN – Edinburgh Office

LAN – London Office

R Router

C Computer S Server
 Internet – the World's biggest WAN?

Autonomous System
(AS)
B
Autonomous System
Autonomous System (AS)
(AS) C
A

Autonomous System
Autonomous System
(AS)
E (AS)
D

Large Scale Structure (Beyond A level)

Performance
The speed of network connections is affected by a number of factors:
• Bandwidth
• Traffic
• Number of devices connected
• Interference and errors
• Latency
Performance - Bandwidth
• Bandwidth is the most important performance factor
• It is a measure of how much information can be transferred through
the network in a fixed time
• Often measured in megabits per second (Mbps) or gigabits per
second (Gbps)
• The higher the bandwidth, the higher the performance
Performance - Traffic
• Bandwidth is 'consumed' when we send information across a
network
• This is often called network 'traffic'
• As we put more and more traffic on a given network, its
performance will decline
• At some point, the performance may collapse, depending on
network protocols
• Applications such as video streaming generate much more traffic
than, say, transferring email
Performance – Number of devices
• Clearly, the more computers or other devices we have connected to
a network, the higher potential traffic levels we have
• But each device also consumes some bandwidth just by being
connected
• And if we have a lot of devices connected to a single network, we
can encounter performance problems when the devices attempt to
access the shared medium at the same time
Performance – Interference and errors
• Many network technologies are vulnerable to electronic
interference
• WiFi technologies are particularly vulnerable – to:
• interfering radio frequency signals (such as mobile phones, baby listeners
and other WiFi networks) and
• radio frequency noise (from switches, fluorescent tubes)
• 'attenuators' or blockers – physical walls, metal structures
• Wired technologies are also vulnerable to a lesser degree
• When errors happen, they can be detected and information
retransmitted, but this reduces performance too.
Performance - Latency
• Finally, some technologies have delays or 'latency' when
transmitting and receiving signals
• These technologies can be high bandwidth, but delays will still
impact performance
• A common example is the use of satellite links – in geostationary
orbit, it takes around ¼ second (250 msec) for a signal to go up to
the satellite and back down to earth
• (Which is why Starlink's satellites are in low earth orbit (LEO) rather
than GEO – delays are around 25 msec, rather than 250 msec)
Peer-to-peer vs Client Server
• In a Peer-to-peer network, each node is equally privileged and acts as
both a supplier and consumer of resources. These shared resources
could be processing power, storage, bandwidth or services such as
printing.
• In a Client-server network, nodes are identified as either Servers
(which provide resources) or Clients (which consume resources).
Peer-to-peer networks
• Many home and small offices networks operate to some degree as
P2P networks, through folder sharing and similar arrangements.
• P2P networking has attracted a reputation for piracy and copyright
infringement through services such as Napster and Web/BitTorrent,
however much of the Internet was P2P in earlier days, and there are
many positive applications of P2P, particularly in trading.
Client-server networks
• Largely as a result of services such as email and the WWW, a
significant amount of internet traffic is now client-server in nature.
• Clients vary in power from very simple (thin) clients, through to
computers of similar capabilities to many servers.
• Servers provide specific services – for example, a web server provides
access to web pages; a file server provides services to upload and
download files. More powerful servers may provide multiple services.
Client-server and peer-to-peer networks

C C C S
C C C C
"Physical" layout
LAN – Birmingham Office
LAN – Edinburgh Office

Client-Server Peer-to-Peer

S C C

"Logical" layout
C C
C C
C
 Client-Server and Peer-to-Peer
Client-Server Peer-to-Peer
• Server has more resources • All computers are of equal status
(processing power, storage) • Each computer provides access to
than clients resources and data
• Server provides clients with • Data is distributed
services
• Computers communicate and
• Data is centralised (and share resources
managed centrally)
• Each computer is responsible for
• Server provides security its own security
• Server is a single point of failure
 Thick and Thin Clients
Thick Clients Thin Clients
• More capable (local storage and • Simple (limited local storage and
processing) processing)
• Can operate without Server to • Send requests to Server and display
some degree the returned results
• More expensive • Dependant on Server
• Cheap
Exercises on OneNote
• Past Paper Questions on the topics covered
• Plus… What can we do for extension? Possibly give them a
complete Paper 1 to work through?
Network Hardware
• Network Interface Controllers and Cards (and Wireless NIC)
• Switches and Hubs
• Routers
• Modems
• Wireless Access Points
• Transmission Media
Network Interface Cards
• Many years ago, personal computers (PCs) were built to work on
their own
• When networks were introduced, Network Interface Cards or NICs
were necessary – you installed a NIC in your PC in order to connect
to the Local Area Network
• NICs are still available, but the vast majority of desktop, laptop,
tablet computers, phones, TVs, and other devices have built-in
network interface controllers
• These built-in controllers perform the same function as Network
Interface Cards
NIC Functions
• Provide the physical connection to a network (wired or wireless)
• Provide a low-level Medium Access Control (MAC) address*
• Implement some network protocols (usually the physical link layer)*

* We will cover these in the next few lessons

Wireless Network Interface Controllers
• Perform the same functions as NICs, but operate wirelessly
• Include an antenna to communicate via radio waves (in the
microwave part of the spectrum)
• As with wired NICs, most WNICs are now built-in to devices rather
than being plug-in cards or connected via USB
Switches
• Network Switches are devices which offer multiple network
connection 'ports'
• Many devices (computers, printers, routers) can be attached to a
single Network Switch
• The Network Switch 'learns' the MAC addresses of the connected
devices, and can 'switch' packets between ports for local traffic
• For packets with remote destinations, the Network Switch can
forward to a router or another switch
• Switches are shared media just like wires or RF, but they are
'smarter' and can direct packets just to the intended destination
Hubs
• Hubs are very similar to switches…
• … but they are much simpler
• Hubs retransmit all packets received on any one port to all other
ports – behaving very much like a physical bus connection
Routers
• Routers connect networks together – allowing traffic (packets) to
flow across the internet
• Each router 'learns' about the networks it is directly connected to
via routing protocols, and some configuration
• Routers can respond to network problems by rapidly re-routing
traffic to avoid networks that are overloaded or out of service.
Routers - continued
• Receives and sends data between two or more networks operating
on the same protocol
• Receives packets from devices or the Internet
• Maintains a table of MAC and IP addresses
• Assigns / allocates private IP addresses to devices on LAN
• Stores / updates / maintains a routing table
• Finds the destination of the packet
• Finds the most efficient path to the destination
• Forwards / routes packets to the destination
 Network Hardware - Router
User Service

Application Application

Transport Transport

Network Network Network Network Network Network

Link Link Link Link Link

Laptop/Phone Router Router Router Router Server

Wireless Routers
• The wireless router that you have in your home is a special kind of
router with some additional functions
• It has a very simple job in many ways – it is effectively connected to
just two networks; your home WiFi network, and your broadband
provider
• So, the routing 'decision' is very simple – unless traffic is going from
one device on your WiFi network to another, then the router just
sends the traffic to the internet via your broadband
• However, your Wireless Router also provides a set of services to
devices on your home WiFi network
 Tablet Firewall DNS
Smart
phone DHCP Routing

Laptop WiFi ISP Autonomous

Wireless connections
Router/AP System (AS)

Desktop
PC Wired connections
Smart
fridge
NAS Smart TV
 Network Hardware - Bridge
User Service

Application Application

Transport Transport

Network Network Network Network Network

Bridge

Link Link Link Link Link

Laptop/Phone Bridge Router Router Router Server

 User
Network Hardware - Gateway Service

Application Application
Relay

Transport Transport Transport Transport

Network Network Network Network Network Network

Link Link Link Link

Router Gateway Router

Bridge, Router, Gateway Comparison
• Bridge
• Connects two networks into a single, logical network.
• Operates at the link layer (on MAC addresses).
• Router
• Directs packets between networks (using the same technology and protocols).
• Operates at the network layer (protocols, tables)
• Gateway
• Directs packets between two dissimilar networks.
• Operates at or above the network layer (often at the application layer),
making it possible for two different networking technologies to work together.
E.g., two email systems.
Modems
• Modems are used where we need to transmit digital data over
analogue networks or media
• Most modems are embedded in devices such as NICs and routers
• The word is a contraction of modulation/demodulation
• With modulation being the process of transforming a digital signal
into an analogue form, and demodulation being the reverse
Transmission Media
• Short-range:
• Twisted-pair cables
• Ethernet cables
• Wireless/RF (WiFi and Bluetooth)
• (Network Switch/Hub)
• Longer-range:
• Fibre-optic
• LoRa/microwave/RF
Ethernet – CSMA/CD
• Ethernet is a large set of standards covering low-level protocols,
connectors, and cables
• A protocol (suite)
• For data transmission over standard / universal wired / cabled network
connections
• Uses Carrier Sense Multiple Access / Collision Detection (CSMA/CD) –
see next slide
• Data is transmitted in frames
• ... each frame has a source and destination address (usually MAC)
• ... and error checking data (so damaged frames can be resent)
Ethernet – CSMA/CD
• A significant part of Ethernet is the CSMA/CD standard:
• Carrier Sense, Multiple Access, Collision Detect
• A computer listens to the communication channel. If the channel is busy, it
will detect a signal (carrier sense)
• ...data is only sent when the channel is free (when no carrier is sensed)
• Because there is more than one computer connected to the same
transmission medium (multiple access)
• ... two computers can start to transmit at the same time, causing a collision
• If a collision happens, the computers abort transmission (collision detect)
• ...and each waits a random amount of time before attempting to resend
Exercises on OneNote
• As before…
Introduction
• Networks and topologies
• LANs and WANs
• Performance
• Client-server and peer-to-peer, thick and thin clients
• Network hardware and Ethernet
• The internet and WWW (and DNS)
• Cloud Computing
• Bus, star, mesh and hybrid topologies
• Wired and wireless networks, protocols, and layers
• Wired and wireless
• Encryption
• Addressing (IP and MAC)
• CIDR and Subnets
• Bitstreaming
 The Internet and the World Wide Web
Internet WWW
• Infrastructure • Built on the Internet
• Network of networks • Huge collection of websites
• Global connections providing data and services
• Supports many services
What is the WWW?
• An information system accessed over the Internet, using TCP
• A collection of interconnected documents and other resources
• Web Browsers and Web Servers
• Core protocol is HTTP
• Core content format is HTML
• Uniform Resource Locators (URLs) for addressing
HyperText Transfer Protocol (http)
• Simple
• Focused on moving HTML documents (and associated documents)
• Uses TCP (and encrypted TCP connections for https)
HyperText Transfer Protocol (http)
Requests

Client
Server
(Browser)
Responses

Request Responses

GET (read) URL 1xx: Information (100 – Continue, 101 – Switching Protocols)

PUT (write) URL + Content 2xx: Successful (200 – OK, 201 - Created)

POST (write) URL + Content 3xx: Redirection (301 – Moved Permanently)

DELETE URL 4xx: Client Error (401 – Unauthorised, 404 – Not Found)

… 5xx: Server Error (500 – Internal Server Error)

Uniform Resource Locator (URL)
https://www.bbc.co.uk/weather/kt17/day2

1 2 3 4

1 "Scheme" – http, https, ftp, mailto,…

2 ":" delimits the Scheme, "//" delimits the 'authority'

3 Host (server)

4 Path
 GET https://www.example.com/path/subpath/doc.html
Client
(Browser)

HTTP/1.1 200 OK
Server
Date: Wed, 27 Nov 2019 15:28:53 GMT
Server: Apache/2.2.14 (Win32)
Last-Modified: Wed, 20 Nov 2019 19:15:56 GMT
Content-Length: 104
Content-Type: text/html
Connection: Closed
<!DOCTYPE html>
<html>
<body>
<h1>Hello World!</h1>
</body>
</html>
URLs – what are they really?
• On a static website, each URL will correspond to a file held by
the website server.
• Any URL which is not known to the website server will cause a 404
error to be generated (although a custom page may be returned)
• On a dynamic website, some URLs may correspond to files,
but others will be more general resource addresses.
• For example: https://www.example.com/members/12734
• This URL may refer to a particular member's record. When the website
server receives this request it may retrieve information from a database
and then create an HTML page which describes/defines the member,
and which contains links which will modify the member's details.
URLs – what are they really?
• In general terms, a URL is an address that we can use to
access a resource held on a web server.
Domain Name System (DNS)
• A way of turning Internet Protocol names (such as www.bbc.co.uk)
into Internet Protocol addresses (such as 171.124.23.12)
• All routing is carried out on the basis of IP addresses
• The process of turning names into addresses is called ‘resolving’
 Examples of domains
Root Level

com tv net uk de Top Level Domains (TLDs)

google co gov google Second Level Domains

mail www lloydsbank bbc hmrc Third Level Domains

www.google.com.
mail.google.com. www news Fourth Level Domains

www.bbc.co.uk.
news.bbc.co.uk.
DNS Resolution
• Resolution can be a recursive process.
• Computers have local DNS resolvers which maintain a small database
of information. If an address is not present (or the information is
stale), then the configured DNS server will be consulted. This server
may in turn consult other DNS servers higher in the 'hierarchy'.
Cloud Computing
• Cloud Computing is the remote provision of servers and services,
accessed over the internet
• Instead of buying and operating servers and services locally,
businesses and individuals ‘rent’ servers and services that are
running in data centres
• Examples include Microsoft Azure and 365, Google Cloud Platform,
IBM Cloud Services, Amazon Web Services
• Cloud Computing may be provided on a 'private' or 'public' basis –
we will discuss this in a moment…
 Cloud – Advantages and Disadvantages
Advantages Disadvantages
Services can be accessed from anywhere in No internet, no service
the world (as long as you have an internet Locations with poor connections cannot make
connection) full use of cloud computing
Temporary interruptions to network access can
cause problems
Easier to share services and information Security concerns
Lower maintenance Higher ongoing costs
Lower upfront costs Reduced control over data and services
Automatic backups/redundancy
Public and Private Cloud Computing
• Public Cloud - Computing services offered by 3rd party provider
over the public Internet
• Public Cloud is open/available to anyone with the appropriate
equipment/software/credentials
• Private Cloud - Computing services offered either over the Internet
or a private internal network
• Only available to select users not the general public
• Private Cloud is a dedicated/bespoke system only accessible
for/from the organisation
Cloud Storage - Pros
• Cloud storage can be free (for small quantities )
• No need for separate (high capacity) storage devices // saves storage
on existing devices
• Can access data from anywhere/any computer with internet access
• Most cloud data services will have in-built backup/disaster recovery
• Security managed by provider, normally very reliable
• Capacity can easily be increased/decreased
• Data can be easily shared
• Multiple people can work on the same data/document/app at the
same time
Cloud Storage - Cons
• Can only access (the cloud) with internet access
• Relies on the providing company for security
• Relies on the providing company for backups
• It can take a long time to upload/download the data
• It can be more expensive in the long term
• There could be a limit to the amount of storage unless paid for
• There could be compatibility/access issues
• There could be issues with the company offering cloud services
Bus, Star and Mesh Topologies
• Networks can use a variety of approaches to connecting computers
and devices (nodes)
• The different arrangements of nodes are called 'topologies' – logical
maps
• There are many different topologies in use, but in A level we
examine three; the bus topology, the star topology and the mesh
topology
• It is possible to combine topologies in a single network – termed
'hybrid' topologies
 Bus Topology Server Device Device Device

• All devices are directly connected to a single, central cable

• Packets can only be sent from one device at any one time Central Cable

• Packets are received by all devices, which need to check

the destination address to see if the packet was sent to Device Device Device Router
them
• Failure of the central cable will affect the whole network
• Failure of a single device/node will not affect the other
nodes (depending on the role of the device)
• Limited security as all nodes see all packets
Internet
• Easy to extend, but performance deteriorates as more
nodes are added
 Star Topology Device
Device
Device

• All devices have their own cable

connecting them to a central hub Server Switch/Hub Router

or switch
• Failure of one device or cable Device Device
does not affect the rest of the Device

network
• Failure of the central hub or
switch will affect the whole
Internet
network
• Performs well under load
 Mesh Topology Device

• Most frequently used in wireless Device Device

networks
• Faster routing of traffic within the
network
• Each device/node can send, receive, and
Device Device Router
relay messages
• Failure of one node does not affect the Device

rest of the network (with one exception!)

• Lower cost (for wireless)
Internet
• Example shown is fully-connected; every
node is connected to every other node.
Some mesh networks are partially-
connected.
Comparison with Client-Server/Peer-to-Peer
• The Star and Mesh topologies are very closely related to the Client-
Server and Peer-to-Peer networks that we looked at in the last
lesson
• The difference is in the 'level' that they operate on. Star and Mesh
are network-level concepts, whereas Client-Server and Peer-to-Peer
are user-level concepts
• It would be possible to run the Client-Server model over either Star
or Mesh networks, and similarly we could operate the Peer-to-Peer
model over a Star or Mesh network
Hybrid Networks
• Hybrid Networks are mixtures of two or more of these topologies
• Advantages include the ability to use appropriate topologies for
different parts of the overall network, which may in turn improve
overall performance
• Disadvantages include the complexity of managing several different
topologies
Exercises on OneNote
• As before, but with some DNS tools work as extension?
Introduction
• Networks and topologies
• LANs and WANs
• Performance
• Client-server and peer-to-peer, thick and thin clients
• Network hardware and Ethernet
• The internet and WWW (and DNS)
• Cloud Computing
• Bus, star, mesh and hybrid topologies
• Wired and wireless networks, protocols, and layers
• Wired and wireless
• Encryption
• Addressing (IP and MAC)
• CIDR and Subnets
• Bitstreaming
Wired and Wireless Connections
• Computers and other devices can be connected to networks either
via wires, or wirelessly
• While wireless connections are very common for phones, tablets,
and laptops, the great majority of connections within the internet
are wired, rather than wireless
• Wireless connections are very convenient for mobile users, but
wired connections are more reliable, and much higher bandwidth
Wired Connections
• Ethernet is the most popular set of standards for wired connections
• The standards cover protocols, connectors, cables, and more
• There are three main types of Ethernet cable:
Twisted pair (Copper) Coaxial (Copper) Fibre Optic
Mainly used LAN LAN/WAN WAN
in:
Bandwidth: Low Medium High
(but not shared)
Distance: Up to 10 km Up to 10 km Up to 100 km
(but usually a few
metres)
Cost: Low Medium High
Wired: Fibre-Optic v Copper
• Fibre optic data is transmitted using light, copper cable through
electrical signals
• Fibre optic has higher bandwidth than copper cable // Fibre optic has
higher transmission rates than copper cable
• Fibre optic has smaller risk of (noise) interference than copper cable
• Fibre optic can be used over longer distances than copper cable before
repeaters are needed
• Fibre optic is much more difficult to hack into than copper cable
• Fibre optic is more prone to damage than copper cable
Wireless Connections
• There are two main standards used in wireless connections – WiFi
and Bluetooth.
• Again, these standards cover a range of topics including protocols
Bluetooth WiFi
Mainly used in: Phones, peripherals, Phones, tablets, laptop
personal devices and desktop computers,
(headphones, activity TVs, Smart Speakers
trackers, smartwatches)
Bandwidth: Low (<1 Mbps) Medium (up to 11 Mbps)
Distance: 5-20 metres Up to 90 metres
(but often a metre or less)
Power requirement: Low Medium
Security: Low Medium*
 Wired and Wireless Connections
Wired Wireless
• Faster connection/higher • Freedom of movement
bandwidth • Easy to expand/add devices
• More reliable/stable connection • No cabling or cabling experience
• More secure required
Satellite V cables
• Satellite … is not fixed to a single location
• … allows access in remote / rural areas

• … high latency / lag / slow to connect (for geostationary, not for LEO
systems such as Starlink)
• … more expensive than wired methods, as we need extra equipment
• … signal is affected by bad weather
• … the transmission speed is slower than fixed line broadband (although
not so much of a problem for LEO)
• … direct line of sight needed (ditto)
Internet Protocol Addresses
• We have already encountered IP addresses, when we looked at DNS
• In the early days of the internet, the structure of IP addresses was
changed rapidly as the scale of the internet ballooned
• Version 4 of the IP address standard was designed to deal with
significant network growth, and we still use IPv4 today in some parts
of the internet
IP Addresses - General
• Public IP addresses are visible to all devices on the internet
• Private IP addresses are only visible to devices within the LAN
• Static IP addresses remain the same each time a device is connected
to the internet
• Dynamic IP addresses are allocated / reallocated each time a device
connects to the internet
IPv4 Addresses
• An IPv4 address consists of 4 bytes, normally written as four denary
numbers, separated by dots:
192.168.4.20
• With this structure, we had 232 addresses available – almost 4.3
billion addresses
• IPv4 worked well for around 20 years, but by the mid 1990s we were
running out of addresses, so we needed another approach
IPv6 Addresses
• The IPv6 address is 16 bytes long (four times the size of IPv4)
fd1c:47a9:1305:0:8a94:64bb:a24f:bd7e
• (We do not mention IPv5 – it did not work very well…)
• The size of an IPv6 addresses gives us many trillions of unique
addresses; more than we can feasibly use
• However, one of the best things about IPv6 is that we can use it
alongside IPv4 (many of your devices will have both an IPv6 and an
IPv4 address)
 IP Addresses – V4 and V6
An IPv4 address contains 4 groups of digits. Each group is represented in 8
bits and the groups are separated by full stops. Size: 32 bits
Example: 192.168.1.1
Example: 127.0.0.1 (special address 'home'/loopback)

An IPv6 address contains 8 groups of digits. Each group is represented in 16

bits. Multiple groups that only contain zeros can be replaced with a :: (double
colon). Size: 128 bits
Example: 2001:0db8:0001:0000:0000:0ab9:C0A8:0102
Example: ::1 (special address 'home'/loopback)
 172 217 169 4

1 0 1 0 1 1 0 0 1 1 0 1 1 0 0 1 1 0 1 0 1 0 0 1 0 0 0 0 0 1 0 0

Class A
0 N N N N N N N H H H H H H H H H H H H H H H H H H H H H H H H

Class B
1 0 N N N N N N N N N N N N N N H H H H H H H H H H H H H H H H

Class C
1 1 0 N N N N N N N N N N N N N N N N N N N N N H H H H H H H H

Old school IPv4. The concept of 'classes' of networks has been replaced by a more flexible approach to using
the address space, coupled with network address translation.
CIDR and Subnets
• Classless Internet Domain Routing (CIDR) did away with the fixed
structure of Class A, B, C
• CIDR introduced the concept of a 'network mask', splitting the IPv4
32-bit address space into a network part and a host part.
• For example, a network mask of 255.255.255.0 (equivalent to a
suffix of /24, meaning 24 bits of network id, 8 bits of host id) would
correspond to the Class C structure
• But CIDR allows us to have any split – for example 255.255.240
(equal to /22, meaning 22 bits of network id, 10 bits of host id).
Subnets
• Reduce amount of traffic in a network // improve network speed //
performance // reduce congestion
• Data stays in its subnet so it does not travel as far
• Improve network security
• .. so that not all devices can access all areas of the network
• Allow for easier maintenance // day-to-day management
• ... because only one subnetwork may need taking down/changing
while the rest of the network can continue
• Allow extension of the network / devices attached
Introduction
• Networks and topologies
• LANs and WANs
• Performance
• Client-server and peer-to-peer, thick and thin clients
• Network hardware and Ethernet
• The internet and WWW (and DNS)
• Cloud Computing
• Bus, star, mesh and hybrid topologies
• Wired and wireless networks, protocols, and layers
• Wired and wireless
• Encryption
• Addressing (IP and MAC)
• CIDR and Subnets
• Bitstreaming
Bitstreaming
• Much of the information sent over the internet does not have strict
timing information – emails, documents, still images
• This information can be sent successfully even at relatively low rates, and
with variable delays
• However, information such as audio and video has very strict timing
information – if we fail to deliver the packets on time, our music or video
will have lags and freezes which the user will be very aware of
• BitTorrent is one of the most commonly used peer-to-peer bitstreaming
technologies, but has been associated with piracy.
• Commercial services such as Netflix and Spotify use other protocols
layered on TCP/IP for bitstreaming.
Bitstreaming – On-Demand/Real-Time
• Bitstreams may be transmitted in real-time (live) or pre-recorded
(on-demand)
Bitstreaming - Compression
• Video and audio information contains significant amounts of
redundancy; data which is repeated, or which changes only
maginally between successive samples
• Compression can therefore be used to significantly reduce the
volume of information that we need to send, even for high-quality
transmissions
Bitstreaming - Buffering
• Packet-switched, public networks such as the internet cannot
provided guaranteed end-to-end bandwidth
• To deal with variations in the rate at which packets are received
bitstreaming uses buffering
• This allows us to play the video/audio stream at a constant rate,
evening out variations in arrival rate
• Only if the buffer is completely emptied will playing of the
video/audio have to pause
Bitstreaming Pros and Cons
• Do not need to wait for the • Streaming depends on internet
whole track/movie to connection – if this is poor or
download unavailable, then the bitstream
• Do not need to store large will be affected
audio/video files • Requires significant bandwidth
• Better piracy protection as it is and a reliable, low latency
more difficult to intercept and connection
copy bitstreams • Some association with
streaming of pirated content
Exercises on OneNote
Older Material – may be useful for A2
 Networking Addressing, Layers and Protocols

IP name + Various protocols – http, https, ftp, smtp, pop, …

Application Application
app
Transmission Control Protocol (TCP)
IP + port Transport* Transport
Internet Protocol (IP)
IP Network** Network Network Network Network

MAC Link Link Link Link Link Link Link Link

*as well as TCP, other protocols such as UDP and DHCP use the Network Layer
**OCR sometimes call the Network layer the "Internet" layer.
OneNote Exercises
• Open Section "1.3 Networks Connections and Protocols"
• Work through the "Lesson 3 Exercises" page
• These lesson slides are available on the "Lesson 3 Slides" page if you
need to refresh your memory
Introduction
• Networks and topologies
• LANs and WANs
• Performance
• Client-server and peer-to-peer
• Network hardware
• The internet and WWW (including DNS)
• Cloud Computing
• Star and mesh topologies
• Wired and wireless networks, protocols, and layers
• Wired and wireless
• Encryption
• Addressing (IP and MAC)
• Standards
• Protocols
• Layers (Remove these and replace with bitstreaming?)
Introduction
• Networks and topologies
• LANs and WANs
• Performance
• Client-server and peer-to-peer
• Network hardware
• The internet and WWW (including DNS)
• Cloud Computing
• Star and mesh topologies
• Wired and wireless networks, protocols, and layers
• Wired and wireless
• Encryption
• Addressing (IP and MAC)
• Standards
• Protocols
• Layers
Standards
• Without Standards, the network systems built by one company
would not work with those built by another company
• In IT1, you use a computer built by Hewlett-Packard, with a
processor built by Intel, running an operating system from
Microsoft, with a NIC from Realtek, and possibly using a browser
built by Google.
• And we have not even mentioned the cables, network switches,
routers and all of the rest of the internet…
• We have Standards for connectors, cables, software, hardware, and
then a particular set of Standards for communication over networks
Protocols – Standards for Communication
• A Protocol is a set of rules which define how two or more devices
communicate
• These rules will define:
• The set of messages that can be used
• Ordering of messages
• Message formats
• What to do when things go wrong
HyperText Transfer Protocol (http)
Request
The set of messages that can be used
Client
Server Ordering of messages
(Browser)
Message formats
Response
What to do when things go wrong

Request Responses

GET (read) URL 1xx: Information (100 – Continue, 101 – Switching Protocols)

PUT (write) URL + Content 2xx: Successful (200 – OK, 201 - Created)

POST (write) URL + Content 3xx: Redirection (301 – Moved Permanently)

DELETE URL 4xx: Client Error (401 – Unauthorised, 404 – Not Found)

… 5xx: Server Error (500 – Internal Server Error)

 Principal Protocols
Protocol Name Description
HTTP/ HyperText Transfer Protocol Web browser to web server protocol
HTTPS (/Secure)
SMTP Simple Mail Transfer Protocol Core email transfer protocol, used by mail
servers
DNS Domain Name System Protocol One part of the Domain Name System is a
Protocol protocol for querying servers
POP * Post Office Protocol Email collection – local email client downloads
all email from the server
IMAP * Internet Messaging Access Email collection – local email client views email,
Protocol but leaves email on server
FTP/SFTP * File Transfer Protocol (/Secure) Upload and download files from a server
TCP Transmission Control Protocol Works over IP to provide information transfer for
most other protocols
IP Internet Protocol Core network protocol of the internet
* Increasingly replaced by web services
Email Protocols
POP/IMAP

SMTP SMTP
SMTP
Web Protocol
Web Server

HTTP

Web Client
Layered Model of Networking - standards

+
HTTP/HTTPS FTP/SFTP SMTP POP IMAP DNS +
+
TCP (UDP)

IP
+
(ATM) WiFi Ethernet Fibre (P2P) (X-25) +
+
 Layered Model of Networking – logical view

Deals with application-specific data – text, graphics,

HTTP, FTP, SMTP, … Application
video, audio, email, files… Works end-to-end
Deals with streams of data, segments and reassembles data, deals with
TCP Transport
missing, out-of-order and damaged data. Works end-to-end.
Deals with packets of data, routes packets between networks, discards
IP Network
damaged packets. Each packet may take a different route. Works end-to-end.
Deals with units (frames, packets) of data, transmits and receives over a
Ethernet, WiFi, … Link single physical medium. Works within a single network.
 Layered Model of Networking - addressing

IP name + Various protocols – http, https, ftp, smtp, pop, …

Application Application
app
Transmission Control Protocol (TCP) or User Datagram Protocol (UDP)
IP + port Transport* Transport
Internet Protocol (IP)
IP Network** Network Network Network Network

MAC… Link Link Link Link Link Link Link Link

*as well as TCP, other protocols such as UDP and DHCP use the Network Layer
**OCR sometimes call the Network layer the "Internet" layer.
Layering - Simplified
Application Application
Data
Application Header Footer

Transport Transport
Data
Transport Header Footer

Network Network
Network Data
Header Footer

Link Link
Link Data
Header Footer
Purpose of Layers in Networks
• Separating network functions into layers makes it easier to focus on
specific concerns
• Layering also supports standardisation – we can combine layers
from different manufacturers
• We can also incorporate new hardware rapidly – we only need to
change one layer of the 'stack'
Ethernet – CSMA/CD
• Ethernet is a large set of standards covering low-level protocols,
connectors, and cables
• A protocol (suite)
• For data transmission over standard / universal wired / cabled network
connections
• Uses Carrier Sense Multiple Access / Collision Detection (CSMA/CD) –
see next slide
• Data is transmitted in frames
• ... each frame has a source and destination address (usually MAC)
• ... and error checking data (so damaged frames can be resent)
Ethernet – CSMA/CD
• A significant part of Ethernet is the CSMA/CD standard:
• Carrier Sense, Multiple Access, Collision Detect
• A computer listens to the communication channel. If the channel is busy, it
will detect a signal (carrier sense)
• ...data is only sent when the channel is free (when no carrier is sensed)
• Because there is more than one computer connected to the same
transmission medium (multiple access)
• ... two computers can start to transmit at the same time, causing a collision
• If a collision happens, the computers abort transmission (collision detect)
• ...and each waits a random amount of time before attempting to resend
OneNote Exercises
• Open Section "1.3 Networks Connections and Protocols"
• Work through the "Lesson 4 Exercises" page
• These lesson slides are available on the "Lesson 4 Slides" page if you
need to refresh your memory
• If you finish everything, start working on Practice Paper 1 in the new
"Practice Papers" section.
Older Material based on OCR A level
From basic communication to networking
• Simple point-to-point communications form the basis for networking
• Access to a shared medium (wired or wireless)
• Need for protocols to govern shared access and communication
• Layered model of networking
• General description
• Multiple models; focus on the TCP/IP layered model
• Packet switching vs circuit switching
• Peer to Peer vs Client-Server
Rules are required to govern access to the shared medium:
How do we ensure that all of the computers have a 'fair' share of access?
How do we resolve conflicts?

Rules are also required to govern the communication between computers:

How do we find the 'address' of the computer we wish to talk to?
How do we hold a 'conversation' with this computer?

A set of rules on a specific subject is called a protocol. TCP, HTTP, IP, FTP, SMTP are all Internet protocols.

Standards are required to ensure that systems can work together, even if they are built by different teams or companies.
A protocol is a particular kind of standard, others standards relate to hardware (e.g., the USB connector) and software
(e.g., the language C#).

Computer Computer Computer Computer Computer Computer Computer Computer

Shared Medium
These standards and protocols apply across networks of networks….

Computer Computer Computer Computer Computer Computer

Router Router

Shared Medium Shared Medium Shared

These standards and protocols apply across networks of networks….
… extending to billions of devices on the internet today.

Computer Computer Computer Computer Computer Computer

Router Router

Shared Medium Shared Medium Sha

Shared Medium Share

 Layered Model of Networking Layer 'N' Interface

Layer 'N' Protocol

Layer 'N' Layer 'N'

Uses services of…

Provides services to… Layer 'N-1' Interface

Layer 'N-1' Protocol

Layer 'N-1' Layer 'N-1'

Uses services of…

Provides services to… Layer 'N-2' Interface

Layer 'N-2' Protocol

Layer 'N-2' Layer 'N-2'
Layered Model of Networking – logical view

Deals with application-specific data – text, graphics, video, audio, email, files…
Application
Works end-to-end
Deals with streams of data, segments and reassembles data, deals with missing, out-of-order and
Transport
damaged data. Works end-to-end.
Deals with packets of data, routes packets between networks, discards damaged packets.
Network
Each packet may take a different route. Works end-to-end.
Deals with units (frames, packets) of data, transmits and receives over a single physical medium.
Link Works within a single network.
More on Protocols…
• A protocol is a set of rules governing communication between two or
more devices.
• TCP and UDP
• Point to point vs point to multipoint; connection-oriented vs connection-free;
reliable vs unreliable
• Concept of streams and packetisation
• Internet protocols managed by the IETF, specified in RFC (Request for
Comments) documents.
• TCP is specified in RFC 793 (with multiple updates in other RFCs).
• Just under 100 pages.
• Defines addressing, messages and their exchange
• Defines security, precedence and robustness approaches.
Layered Model of Networking

Application Application

Transport Transport

Network Network Network Network Network

Link Link Link Link Link Link Link Link

 Layered Model of Networking - addressing

IP name + Various protocols – http, https, ftp, smtp, pop, …

Application Application
app
Transmission Control Protocol (TCP) or User Datagram Protocol (UDP)
IP + port Transport* Transport
Internet Protocol (IP)
IP Network** Network Network Network Network

MAC… Link Link Link Link Link Link Link Link

*as well as TCP, other protocols such as UDP and DHCP use the Network Layer
**OCR sometimes call the Network layer the "Internet" layer.
Layered Model of Networking - standards

+
http ftp smtp pop dns tftp rtp +
+
TCP UDP

IP
+
ATM WiFi Ethernet Fibre P2P X-25 +
+
Peer-to-peer vs Client Server
• In a Peer-to-peer network, each node is equally privileged and acts as
both a supplier and consumer of resources. These shared resources
could be processing power, storage, bandwidth or services such as
printing.
• In a Client-server network, nodes are identified as either Servers
(which provide resources) or Clients (which consume resources).
Peer-to-peer networks
• Many home and small offices networks operate to some degree as
P2P networks, through folder sharing and similar arrangements.
• P2P networking has attracted a reputation for piracy and copyright
infringement through services such as Napster and Web/BitTorrent,
however much of the Internet was P2P in earlier days, and there are
many positive applications of P2P, particularly in trading.
Client-server networks
• Largely as a result of services such as email and the WWW, a
significant amount of internet traffic is now client-server in nature.
• Clients vary in power from very simple (thin) clients, through to
computers of similar capabilities to many servers.
Client-server networks
• Questions?
LAN/WAN
• Local Area Network – all of the connected devices are within a single,
relatively small area. Often associated with an organisation such as a
small business, a school, or a home. Usually owned and administered
by a single organisation.
• Wide Area Network – a network of networks covering a large
geographical area, possibly across countries and continents. Owned
and administered by several cooperating organisations.
Layering
Application Application
Data
Application Header Footer

Transport Transport
Data
Transport Header Footer

Network Network
Network Data
Header Footer

Link Link
Link Data
Header Footer
 IPv4 Packet Structure

10 bytes 2 bytes 4 bytes 4 bytes 4 bytes Variable number of bytes

Header Source IP Destination

Header Options Data
Checksum Address IP Address

4 bits 4 bits 6 bits 2 bits 16 bits 16 bits 3 bits 13 bits 8 bits 8 bits

Header Type of Total Fragment Fragment Time to

Version ECN Flags Protocol
Length Service Length Id Offset Live

1 byte 1 byte 2 bytes 2 bytes 2 bytes 1 byte 1 byte

Note – header protected by Checksum, everything else unprotected.

Role of TTL.
 172 217 169 4

1 0 1 0 1 1 0 0 1 1 0 1 1 0 0 1 1 0 1 0 1 0 0 1 0 0 0 0 0 1 0 0
 172 217 169 4

1 0 1 0 1 1 0 0 1 1 0 1 1 0 0 1 1 0 1 0 1 0 0 1 0 0 0 0 0 1 0 0

Class A
0 N N N N N N N H H H H H H H H H H H H H H H H H H H H H H H H

Class B
1 0 N N N N N N N N N N N N N N H H H H H H H H H H H H H H H H

Class C
1 1 0 N N N N N N N N N N N N N N N N N N N N N H H H H H H H H

Old school IPv4. The concept of 'classes' of networks has been replaced by a more flexible approach to using
the address space, coupled with network address translation.
IPv6
• Introduced to tackle the problem of IPv4 address exhaustion
• Uses a 16 byte address in comparison to IPv4's 4 byte address
• In theory, allows an address space of 2 to the power 96.
• However, the address space is used inefficiently (intentionally) to
make it easier to route packets
• For example: 2001:0000:4136:e378:80000:63bf:3fff:fdd2
• (Actually contains an IPv4 address encoded to IPv6 – 65.54.227.120)
• facebook.com = 2a03:2880:f164:81:face:b00c:0:25de
 Organisationally Unique Identifier (OUI) NIC Specific Identifier

Byte 1 Byte 2 Byte 3 Byte 4 Byte 5 Byte 6

b b b b b b b b
7 6 5 4 3 2 1 0

0: Unicast
1: Multicast

0: Globally unique
1: Locally managed
4 bytes 6 bytes 6 bytes 6 bytes 2 bytes Variable number of bytes 4 bytes

Access Point Transmitter Receiver Sequence

Header Data CRC
MAC MAC MAC Number

(Slightly simplified) WiFi Link-Level Frame. Based on 802.11

 Network Hardware - Router
User Service

Application Application

Transport Transport

Network Network Network Network Network Network

Link Link Link Link Link

Laptop/Phone Router Router Router Router Server

 User Service

Application Application

Transport Transport

Network Network Network Network Network Network

Link Link Link Link Link

Laptop/Phone Router Router Router Router Server

 Service
User

Application Applications and the Application

World Wide Web

The Internet
Transport Transport

Network Network Network Network Network Network

Link Link Link Link Link

Laptop/Phone Router Router Router Router Server

 Network Hardware - Bridge
User Service

Application Application

Transport Transport

Network Network Network Network Network

Bridge

Link Link Link Link Link

Laptop/Phone Bridge Router Router Router Server

 User
Network Hardware - Gateway Service

Application Application
Relay

Transport Transport Transport Transport

Network Network Network Network Network Network

Link Link Link Link

Router Gateway Router

Networking Devices
• Router
• Directs packets between networks (using the same technology and protocols).
• Operates at the network layer (protocols, tables)
• Bridge
• Connects two networks into a single, logical network.
• Operates at the link layer (on MAC addresses).
• Gateway
• Directs packets between two dissimilar networks.
• Operates at or above the network layer (often at the application layer),
making it possible for two different networking technologies to work together.
E.g., two email systems.
Networking Devices (continued)
• Firewall
• Filters packets based on source and destination addresses, ports, protocols,
and content. We will cover this in more detail when we look at network
security. Operates at and above the network layer.
• Repeater
• Boosts the signal to extend the range of a physical (sub link layer) network.
• Network Address Translator/Port Mapper (part of a router)
• Enables private networks (usually 192.168.1.x in IP terms) to connect to the
public internet. Operates at the network layer.
 Tablet Firewall DNS
Smart
phone DHCP Routing

Laptop WiFi ISP Autonomous

Wireless connections
Router/AP System (AS)

Desktop
PC Wired connections
Smart
fridge
NAS Smart TV
Protocol Example: DHCP
• How do the devices on a network get their IP addresses (and other
configuration)?
• Impossible to configure by hand on an internet scale
• Dynamic Host Configuration Protocol (DHCP)
• Consists of four messages exchanged between a 'client' and 'server'
DHCP DISCOVER
Broadcast by a client which needs an address/has just booted
Broadcast at IP level and at MAC level.
Header:
• IP Source = 0.0.0.0
• IP Destination = 255.255.255.255
• MAC Source = device’s MAC address
• MAC Destination = FF:FF:FF:FF:FF:FF
DHCP OFFER
Sent by a server (such as your WiFi router), offering to supply an IP address and
other configuration.
Header:
• IP Source = DHCP Server’s IP Address
• IP Destination = 255.255.255.255
• MAC Source = DHCP Server’s MAC address
• MAC Destination = device’s MAC address

Note that there may be a number of DHCP Servers offering configuration

information. At this point, the device will have to pick one of the offers, if there is
more than one.
DHCP REQUEST
Broadcast by the client
Header:
• IP Source = 0.0.0.0
• IP Destination = 255.255.255.255
• MAC Source = device’s MAC address
• MAC Destination = FF:FF:FF:FF:FF:FF

Note that the request, although intended for just one server, is actually broadcast at the IP
and MAC level, so that any other servers that made offers can withdraw those offers. The
offer messages all have unique identifiers; the request message includes the identifier of
the offer it is accepting, so that other servers know that their offer has been declined.
DHCP ACKNOWLEDGE
Sent by the server
Header:
• IP Source = DHCP Server’s IP Address
• IP Destination = 255.255.255.255
• MAC Source = DHCP Server’s MAC address
• MAC Destination = device’s MAC address

Contains a set of configuration information – see next slide

DHCP ACKNOWLEDGE - Contents
• IP Address – essential
• Subnet mask
• Default Router IP address
• Lease Time
• DNS Servers

(There are a number of other codes defined in the DHCP standard)

Clients can renew their ‘leases’, and so hold on to their IP addresses.
They can also release a lease.
Domain Name System (DNS)
• The DNS maps IP names such as www.google.com to IP addresses
such as 172.217.169.4.
• DNS manages associated information on internet domains, including:
• The SMTP mail servers for a domain
• The DNS servers for a domain
• Aliases for a domain
• "Start of Authority" information about the domain's DNS data itself
• More than a protocol – DNS is a database of information plus the
protocols to maintain the database and to respond to queries.
• Extended (e.g., SPF) and misused (as a database).
Domain Name System (DNS)
• Hierarchical approach – see diagram.
• . as the root.
• Fully Qualified Domain Names – full, plus the dot at the end
(technically).
• Changes in the hierarchy – new TLDs. Move away from the (only ever
partially-followed) geographical approach.
• (Use of DHCP to set the initial DNS server information for a host.)
 Examples of domains
Root Level

com tv net uk de Top Level Domains (TLDs)

google co gov google Second Level Domains

mail www lloydsbank bbc hmrc Third Level Domains

www.google.com.
mail.google.com. www news Fourth Level Domains

www.bbc.co.uk.
news.bbc.co.uk.
DNS Resolution
• Resolving names (e.g., news.bbc.co.uk) to IP addresses.
• Resolution is/can be a recursive process.
• Computers have local DNS resolvers which maintain a small database
of information. If an address is not present (or the information is
stale), then the configured DNS server will be consulted. This server
may in turn consult other DNS servers higher in the 'hierarchy'.
 Internet
Society Manages namespaces and numerical spaces
Runs the root name servers
IETF Internet Corporation for Manages .org
Assigned Names and Manages IPv4 and IPv6 address allocations
Numbers Manages internet protocol identifiers
IAB (ICANN)
Delegates some responsibilities to the five
Regional Internet Registries (RIRs)

Manage regional
allocations
AfriNIC ARIN APNIC LACNIC RIPE NCC Delegate some
responsibilities to (1000s
of) Local Internet
Registries, including
Internet Service
Providers, DNS registrars

NOMINET Manages .uk, .wales, .lon

NOMINET don, .bbc …
NOMINET
NOMINET
Manages UK registrars
DNS
• Questions?
Network Security
• Threat overview
• Counter-measures
 Network Threats
Intrusion Malware Man in the Middle Denial of Service Social engineering Spying

Hacking/cracking Worms Impersonation Flooding Phishing Tracking

SQL injection Viruses Distributed DOS Spear phishing Key logging

Buffer/heap/stack Trojans Wiretapping

overflow
Ransomware Port scanning

Rogue security Web beacons

software
('scareware')
Rootkits

People are both the largest threat and the largest vulnerability
Worms
• Malicious software programs that are standalone, and can self-
replicate (often via email).
• They spread via security failures within networks of computers.
• Worms cause damage by themselves (slowing computers and
networks, sometimes to large degrees), but can also carry 'payloads'
such as ransomware, keyloggers, or programs designed to steal data.
 Viruses
• Malicious software programs that rely on other programs in order to
spread.
• For example, a computer game could be 'infected' with a virus; when the
computer game runs, the virus also runs and can then infect other programs.
• Viruses can also be attached to 'macro' files used by common office tools.
• Viruses are created to self-replicate (copying themselves into other
host programs) in order to cause widespread damage.
Trojans
• Malicious software programs that masquerade as useful games or
utilities.
• Once downloaded and run, the Trojan releases its malicious payload,
which will then attempt to steal data, encrypt files, run malicious
services, or make the computer vulnerable to other attacks.
• Trojans do not attempt to replicate themselves further. They rely on
user actions to spread.
Security Measures
• Firewalls
• Proxies
• Encryption
• Anti-virus software
• Regular updates – operating system and virus protection
• Security testing
 Rules
Accept, reject, drop
Firewall Server Layer 3 device, with additional checks
DMZs

Block access to undesirable Prevent unauthorised access and

servers and services intrusion attempts

Firewall Server
Internet

Internal
Network NIC NIC
 Privacy
Monitoring/filtering
Proxy Server Caching
"Application-level NAT"

HTTP GET 1 2 HTTP GET

HTTP RESPONSE 4 3 HTTP RESPONSE

Proxy Server

Internet

Internal
Network
 Virtual Private Networks
C
C

LAN C

WAN

WAN
WAN

S S
C C
WAN
LAN LAN
C C
C C C
C Encrypted
connection
 "Onion" Routing
C
C
S
LAN C

WAN

WAN
WAN

S S
C C
WAN
LAN LAN
C C
C C C
C
Circuit Switching vs Packet Switching
• Overview
Circuit Switching

A B
Circuit Switching

A B
Packet Switching

A 32
1 B
Circuit Switching vs Packet Switching
• Circuit switching was the basis for many early communications networks.
• A dedicated circuit is established between the two communicating
devices (this circuit may be real or virtual).
• As long as the devices need the circuit, it is reserved for their use.
• Circuit switching gives a guaranteed service – speed, order of delivery,
quality.
• Circuit-switching is wasteful however, as fixed resources are allocated to
a circuit whether information is flowing or not. Bandwidth is typically
fixed, and setting up and tearing down circuits adds some overhead.
Circuit switching also degrades poorly, denying any new connections.
Circuit Switching vs Packet Switching
• Packet switching works by separating a stream of information into
packets, and then routing each of these packets individually between
two communicating devices.
• Packet switching is highly efficient, and degrades in a controlled
fashion.
• The big disadvantages of packet switching are that the order of
delivery is not guaranteed, and that quality (particularly latency) can
be highly variable. Packets can be delayed or even lost.
Circuit Switching vs Packet Switching
• Although circuit switching has a number of advantages, packet
switching is the dominant approach.
• Even for time-critical services such as audio (interactive) and video
(higher bandwidth), packet switching can provide a suitable service
with some additional techniques.
Circuit Switching vs Packet Switching
• Questions?
• TCP?
Internet – Large Scale Structure

Autonomous System
(AS)
B
Autonomous System
Autonomous System (AS)
(AS) C
A

Autonomous System
Autonomous System
(AS)
E (AS)
D
 Organisation Peering Organisation Peering Organisation
"Backbone" Mainly IPv6
(Multiple AS) (Multiple AS) (Multiple AS)

Tier 1 - Global Transit Transit Transit Transit Transit Transit

Autonomous System Peering Peering Peering

Autonomous System Autonomous System Peering Autonomous System
(AS) (AS) (AS) (AS)

Tier 2 - Regional Transit Transit Transit Transit

Autonomous System Peering Autonomous System Peering Autonomous System

(AS) (AS) (AS)

Tier 1 - Local