In the days before networks were commonly available, the world of computing was a little dull. Nowadays, computers are fastened together by default. In this section, we will learn some of the basic methods by which computers are fastened together and the characteristics of how they operate. We are learning ...
So that we can ...
Before we start, it's time to delve into the past and watch a pivotal video - "A Communications Primer". Produced by Charles and Ray Eames in 1953, it was intended to introduce the concepts of communication to designers and architects of the time to encourage them to incorporate ideas of communication into their work. Definitely worth watching as it introduces and explains lots of the concepts that we are learning about here.
22 minutes of retro joy :)
Networked devices can be arranged in a Local Area Network (LAN) where all the devices are in one geographical area or in a Wide Area Network (WAN) where the devices are spread out over a wider geographical area. The advantages of networking (sharing hardware, software and other resources, communication, backup, user administration) far outweigh the disadvantages (dependency, bottlenecks, security) in most cases. Networking enable interoperability but requires tight control over communications protocols.
First, some definitions ...
Data is transmitted in binary form, though not like water in a pipe
We can't physically send '1's and '0's down a wire, so how is data transmitted? Logically, there has to be some way of representing '1's and '0's on a wire - voltage levels on a DC line or signal amplitudes on an AC line are often used for this. In so called 'uni-polar encoding' ...
The voltage 'travels' from the sender to the receiver at around 230,000,000m/s; waves travel at the speed of light - so essentially, if I apply a high voltage to or an oscillation on a wire, it is sensed immediately at the other end. We need to get away from the idea that data 'travels' down wires like water through a pipe; more correctly, it is applied by the sender and sensed by the receiver. The 'pressure' is felt immediately it is applied :
However, since efficient transmission of data over moderate distances using DC voltages is not practical (and certainly not possible in unguided media or optical fibre), digital transmission usually requires the use of frequency or a combination of amplitude and phase modulation techniques and the presence of a MODEM (modulator / demodulator) for transmission along an analogue carrier.
Networking topologies
In the context of networking the term topology means "the logical shape, layout, configuration or structure of the connections that join devices to a network." There are two common network topologies - bus and star.
Each network interface adapter is identified by a MAC or Media Access Control address. The MAC address is unique amongst all NICs ever manufactured. Crazy. They look like this - 00:15:E9:2B:99:3C (48 bit).
In a bus network, the devices are connected to a common, high capacity wire called a bus. Devices on the bus are connected using a specialist device called a tap or, sometimes, a vampire tap. In a star network, the devices are connected to a central device like a hub or a switch using a twisted pair cable. Can you see why it's called 'Twisted Pair Cable'?
Hubs and switches First, a bit of news from 2015 ... Town Crier announces birth of Royal baby girl (0:48) What has this got to do with networking? Star networks always have a device at the centre of the star with multiple network ports. The choice is either a network hub or a network switch. Very few networks use hubs these days, for a very special reason ...
Wireless networks (WLAN and WWAN)
There are two commonly used wireless network types - WLAN or Wireless Local Area Network and WWAN or Wireless Wide Area Network. WLANs are generally used in domestic or business settings whereas WWAN is usually used to connect to the internet from a mobile device - a so called 'data connection'.
Watch & Learn - Introduction to Wireless Networking (4:30) Tip 1 : The Wireless LAN uses a unique 32-bit code (name) called a Service Set Identifier or SSID to uniquely identify the network to the clients. The client network card is authorised to access the WLAN with a particular SSID.
Tip 2 : Wireless security helps to prevent unauthorised access to your network and encrypts network traffic
Bluetooth
Physical (and software) networking protocols
First, just to get a grip on exactly how many different networking protocols there are (and this is probably not a complete list), have a look at the document called Network Communication Protocols in the lesson resources or click on the image below (which opens the same document). You will probably have to download the PDF to view it correctly. How many protocols do you recognise? This list covers physical and software protocols.
Click to open document - you might have to download
it to view it properly because the writing is so small ... Communication protocols cover such diverse areas as ...
One of the most common networking protocols (that we've briefly met before) is Ethernet. The following three videos introduce Ethernet and how it works. Whilst you do not need to know this for the examination, it gives you an idea of how data communication occurs on a shared medium - important for the next section on collision detection. Don't worry about references to the TCP/IP protocol stack - we'll meet this again at A Level.
Problems of communication over a shared medium (3:30)
How to build an Ethernet frame (5:21)
Minimal length of an Ethernet frame (4:22)
Essential goodies, yeah! The ethernet frame is composed of ...
Networking collisions and CSMA/CD and CSMA/CA A collision on a bus network. How about playing the 'number shouting game' to help you understand! During this section, we will learn about ...
When transmitting over a shared medium (bus networks and WiFi), there is always a possibility of interference of the signals which are transmitted by different devices. In order to remove the chances of these 'collisions' occurring, devices employ algorithms to (D)etect or (A)void them. First, watch the video about CSMA/CD.
Ethernet collision detection algorithm (7:51)
CSMA/CD is only used on wired networks. On a wireless shared medium, there is no way that the transmitting station can 'hear' interference and detect a collision. The wireless adapter can only monitor the wireless channel for traffic, waiting for a 'gap' before transmitting.
Network segmentation Sometimes, bus networks are segmented in order to reduce collisions and increase performance. Segmenting a bus network using a network bridge Networks are segmented based on IP Address Ranges (we'll not meet this concept until next year). The network bridge acts as a repeater and connects one part of the network to another. The bridge 'knows' which nodes are connected to which port (through their IP address or MAC address) and examines the data packets and routes them to the correct segment of the network. You can also connect network segments through a router. Networks are often segmented to aid administration.
Data Transfer Rate (A Level Only) The first thing to realise is that network data transfer rates are measured in multiples of bits per second whereas data storage is (normally) measured in multiples of bytes.
Of course, when the number of bit gets larger, we have to use prefixes for the data transfer ...
It's a very common mistake to quote networks speeds in multiples of BYTES per second but this is WRONG!
Typical data transfer rates which you might be familiar with include ...
However, these are theoretical limits and in practice, very hard to achieve due attenuation / distortion (see below).
Logical and physical network topology It's often easier to represent complex networks in a logical topology rather than a physical one. Logical topology removes the networking hardware from the network and merely shows the logical connections between the workstations and servers in a network.
In a peer-to-peer network, all networked devices exist at the same hierarchy. They can exist as both client (consuming data) and server (providing data) depending on the task. In a client-server network, there is one networked devices which acts above the other devices - acting as a controller and a server of data. The client devices merely consume.
When you are setting up a client / server network, attention must be paid to where the bulk of the processing will be carried out. If the client is expected to handle many different operations and only rely on the server for user administration and centralised storage, then a thick client system might be appropriate. If the clients will only be expected to carry out simple, narrow operations, with the bulk of the processing occurring on the server, then a thin client system might be appropriate. Thin clients can be physical machines with limited or no secondary storage or can run as virtual machines / software machines.
Serial data transmission
Serial transmission methods have been around since the the early days of the telegraph. This video (from the 1930s) shows how a photograph is transmitted through a telephone line (from archive.org).
Serial transmission is analogous to a single track road where the 'data' is sent and received in sequence.
Simplex, half and full duplex
These terms refer to the direction that data communication occurs in.
Parallel data transmission
Another arty video - this time a proper art installation called "A Parallel Image" by Gebhard Sengmüller. If you are interested in finding out more, click on the link or look at the brochure in the lesson resources.
A Parallel Image (3:13) In parallel data transmission, data is sent simultaneously along parallel wires. Parallel communication is only used for short range communication for a variety of reasons. In fact, it's not used often these days, even inside computers .(PATA > SATA).
Synchronous and asynchronous transmission
In synchronous mode, sending and receiving computers synchronise a clock during transmission using a timing signal. No demarcation in the data packets are required.
In asynchronous transmission, the transmitter and receiver are not kept synchronised. They are synchronised temporarily for the duration of each transmission. Asynchronous transmission is commonly used on low speed serial links which transmit individual characters separately rather than in groups as packets / frames of data. For example, it can be used on the serial link between a keyboard and a computer. The transmitter and receiver have separate clocks that run at (approximately) the same speed. Each character that is transmitted is sent with a start bit, stop bit(s) and parity bit which are used to ensure that the data is received correctly.
Baud rate
Baud rate (unit Hz) is the rate at which signalling events occur in a communications channel. One baud is equivalent to one signalling event per second. For a modem, this could correspond to a change in frequency or phase whereas for a line code it could be a voltage pulse.
Bit rate
Bit rate (unit bits per second or bps) is the rate of bit transmission along a transmission media. It must be at least the same as the baud rate but can be higher ...
The effects of attenuation and distortion make it difficult for multiple voltage signals like the ones above to be received reliably so techniques such as frequency and amplitude / phase modulation are often used instead.
Bandwidth
In electronics, bandwidth is the range of signal frequencies that a particular medium can transmit without significant reduction in signal strength. Essentially, this range of frequencies places a limit on the maximum bit rate (number of bits of data that can be transmitted per second) - a high bitrate signal can only be effectively transmitted over a high bandwidth channel. 60 Symbols Video - great explanation of bandwidth vs data rate (11:38)
Click to enlarge
Nyquist also determined that "the maximum theoretical bit rate for a transmission medium with bandwidth B Hertz is 2B bits per second, when binary data is being transmitted". So, for instance, an 8kbps bitrate signal (8000bps) can be be effectively transmitted over a channel with a 4kHz bandwidth. Think back to the work we did on sound sampling. You have to take a binary sample of the height of a wave at at least twice the frequency of the wave in order to faithfully reproduce it. This is the opposite - one wavelength of the carrier signal holds two binary samples, so the wave carries data at twice it's frequency. See?
Nyquist theorem (20:11)
In computing, however, the term bandwidth is synonymous with bit rate.
Latency
Latency is the time delay between some event occurring and it's effects being felt. In communications terms, you only really notice this when you are watching satellite broadcasts from wartorn countries or in the old days when you were speaking to your friend in Australia using analogue phone technology.
Click to enlarge
Attenuation and distortion
Signals can be ...
How about these?
END OF TOPIC ASSESSMENT
|