CS - Introduction to computer networks & Network Interface Layer

A network is formed when you connect things.

Most experts agree that a computer network is formed when two or more computers communicate.

• Wide Area Network (WAN)

WANs span large geographical areas, typically countries or continents.

• Metropolitan Area Network (MAN)

MANs connect sites around a town or city.

• Local Area Network (LAN)

LANs connect computers within a building or site.

• Campus Area Network (CAN)

Somewhere between a LAN and a MAN, this describes connections over a village-sized area (usually restricted to larger universities).

• Tiny Area Network (TAN)

Used by some people to describe small networks within a house or small office.

A protocol is a set of rules that establish how to do something.

A communication protocol is a set of rules that governs the data communications and enables two entities to exchange information in an orderly manner and free of errors. Without a protocol, two devices may be connected but not communicating.

Each layer:

• Solve a specific problem of communication.
• It sets its own rules or protocols.
• It offers its services to the next higher layer.
• Use the services offered by the layer immediately below.
• Know what makes the bottom layer but not how.

In the source of the communication, each layer:

• Recieves data from the upper layer.
• Encapsulates data.
• Add to the data received some control information that the layer needs to be able to work.
• The resulting product passes until the last layer, where transmission occurs.

A protocol stack is a set of protocols used in a system or communications network, each one of these protocols associated with a particular layer of the system.

The design in levels makes it easier to solve problems by making the global problem into smaller problems, and it's easier to modify any protocol without affecting the others.

The main problem that we intend to solve is the ability to communicate two applications regardless of where (on which computer) are located.

It can be broken down to these 4 layers:

• Network Interface Layer

It enables the exchange of information between two computers that are connected to the same physical medium (belong to the same network).

Protocol: depends directly on the concrete physical environment that is used in the network. Examples: Ethernet, 802.11 or WiFi, or FDDI fiber optic ring.

Most important control information: source and destination MAC addresses.

• Internet Layer

It enables the exchange of information between two hosts who do not share the same physical medium. This level will be in charge of carrying the information of the computer where is the source application, to the computer where the destination application is.

Protocol: IP protocol (as of today, IPv4 is used, but IPv6 will be used in the near future).

Most important control information: source and destination IPs.

• Transport Layer

It enables the exchange of information between applications, regardless of the computer they are. This level provides the information to the destination application within that computer.

Protocol: TCP and UDP.

Most important control information: ports of origin and destination.

• Application Layer

It sets rules that the two applications will follow to communicate, and it sets the sequence of messages that the two applications involved will exchange.

Protocol: depends on the applications that are going to communicate.

Most important control information: each protocol uses its own control information.

At this level is set:

• The physical media that will be used to transmit information.
• How to work with each physical media (how it encodes and transmits the information).
• How must be the devices, cables, antennas, etc... to be used to connect your computer to the physical environment.
• What problems or errors may appear in each of the physical media and how to avoid or solve such problems or errors.
• The position where the different hosts will be placed.
• How will be coordinated the communication between two computers through the same physical environment, etc...

The network interface level is the only one to know how information is transmitted through the physical environment it's connected with.

The remaining layers doesn't know anything about the medium used to transmit information. They are only aware of the fact that all data that they send to the network interface layer will be transmitted through the medium, regardless of the used medium.

All mechanisms for encoding information, so that it can be transmitted through a particular physical environment, are implemented in the network card itself. That's why there are different network cards depending on the medium to which we are connecting: optical fiber, coaxial cable, wireless, etc...

The data transmission can be performed in both directions from one end to another.

There are 3 types of data transmission:

• Simplex

Transmission occurs in one direction only.

If you want to transmit to the contrary, it will be necessary to put another wire.

• Half-duplex

Transmission can take place in both directions, but not simultaneously.

You must use control signal to indicate whether the medium is busy or is free to transmit.

• Full-duplex

Transmission can occur in both directions at the same time on the same cable.

The Bitrate is the number of bits that can be transmitted through the medium in one second.

This measure is expressed in bits per second or bps, with all its multiples: Kbps, Mbps, Gbps, etc..., having between each of them a difference of 1000 units.

Each physical medium has associated a bit rate limit. The bit rate limit will depend also on the physical medium quality.

Through physical media may transmit always slower than the speed limit for that medium, but never faster.

• Guided media

They use a solid medium (cable) for data transmission.

It transmits electrical impulses or lighting. The bits are converted into
network card and become specific lighting and electrical signals that
are determines by the rpotocol that implements the network.

• Unguided media

They are wireless media.

It's based on the propagation of electromagnetic waves through the space.

A pair of plain copper wires is a poor medium for electronic signals, and is therefore rarely used in computer networking.

Electro-Magnetic Interference (EMI): when a signal current passes through a wire or circuit, it creates small electrical and magnetic fields. These fields can affect signal currents in other wires, or other circuits, adding electrical noise to the signals, degrading them.

Long, plain copper wires are easily affected by EMI — hence they should not be used to carry computer signals more than a few meters.

They are normally used as a telephone cable to transmit analog voice.

• Connectors (example)

In co-ax, a copper wire is surrounded by a copper screen.

This screen (usually a fine braid or mesh of copper filaments) is connected to signal ground and it absorbs EMI, protecting the signal wire from electrical noise. Screen and signal wire are separated by plastic insulation.

Co-ax offers high bandwidth, low noise and attenuation, and is excellent for computer networking over hundreds of meters. However, it is expensive to buy and install.

A range of different co-ax cables is available, identified by codes. Popular types used for computer networking include RG-8 and RG-58.

• Bandwidth

The closeness of the screen and signal wire creates an effect called capacitance.

This limits the range of frequencies (bandwidth) that may be carried through the cable.

• Attenuation

This capacitance and electrical resistance of the copper creates an impedance to signals, measured in ohms per metre (Ohm/m), reducing the strength of the signal over a long distance.

This reduction is called attenuation and it's measured in decibels (dB).

• Connectors

RG-8 'thick' co-axial cable is terminated with N connectors. These bulky connectors are also used with radio equipment.

RG-58 'thin' co-ax is usually terminated with BNC connectors. These are also associated with oscilloscopes, older video recorders and radio equipment.

If pairs of plain copper wires are wrapped around each other, this helps cancel-out EMI.

Twisted pair cables offer good bandwidth but are not as good as co-axial cable.

However, this kind of cabling is very cheap to buy and install.

STP costs more but offers better electrical characteristics.

The most popular kind of UTO used for networking is Category 5e ("Cat5") cable.

• Connectors

The standard connector used with Cat5 cable is the RJ45.

After sorting individual wires into appropiate channels, the connector is crimped onto the wires.

It's comprised of a wire made of glass (or plastic) fibers.

Each filament has a core of glass fiber with a high refractive index which is surrounded by a layer of similar material (caddling) but with a lower refractive index.

The whole set is protected by other insulating layer (coating) that absorbs the light.

Optical fiber transmits signals as pulses of light along a flexible glass tube. It does not use electricity, except to power the transmitting and receiving circuitry at both ends.

Fiber-optic wires have exceptional bandwidth, zero EMI and very low attenuation, and can carry signals for miles. However, they are expensive to buy and install.

Radio waves are electromagnetic waves whose wavelength is greater than 30cm.

They are able to travel long distances, and can pass through solid materials such as walls or buildings.

Radio waves are used for multicast communications, such as radio, television and paging systems.

Microwaves are used for unicast communication such as cellular telephones, satellite networks, etc.

These waves travel in a straight line.

Due to the curvature of the Earth, the distance between two repeaters should not exceed about 80km away.

Infrared are directional electromagnetic waves that are unable to pass through solid objects.

TV remote control and Wii remote control.

Infrared signals can be used for short-range communication in a closed area using line-of-sight propagation.