Design  | SFC  |  NOS  |  Code |  Diary  | WWW |  Essay |  Cisco | [Home]



[Bill's Home Page]
[ Old page]


Subscribe now to the Institute of Network Agents, by taking the following exam.
Describe, in great detail, how Class X IP addresses are used to access WWW sites which have objectionable content. Give examples of these domains which you have accessed, giving exact times and their content. Where on your computer might these files be stored, and why are the data packets invisible to the Internet?
Explain the technology that allows gigabit per second speeds down a standard piece of door bell cable. How is it possible to transmit data, and also front door bell signals, along the same cable?
Carefully explain how the door bell signals could be decoded by any host on the Internet, and then how these could be used to open the door (Note: do not use the example of a person actually getting up to go to the door). Also, what happens when the user uses the wrong voltage rating for the front door bell battery?
Show how it is possible to avoid network collision in Ethernet with users waving flags when they are about to send data. How is priority of transmission built into this system, and how is it difficult for a user to predict when they are about to receive data. Also, how might different coloured flags be used? Finally, if different coloured t-shirts are used instead, what effect does this have on the latency of the data packet. that coloured flags would not.
To the nearest thousand, how many users currently connect to the Internet (1pm, Wednesday, 24 January 2001)? Determine, using complex mathematical formulas, the percentage of people who, at this current hour, connect to the Internet in each of the countries around the world. If you have time illustrate these using coloured pens/pencils (Note: marks will be given for neatness, and also the usage of colours. If possible use light colours for areas where there is a light connectivity to the Internet, and dark colours for the rest. You can ignore Antarctica.)
Explain, in great detail, why it is not possible to get connected to your free unmetered access ISP after 6pm at night. Also describe how they will cut you off, just as your accessing important information. What is the protocol that they use so that you cannot connect back to the Internet for at least eight hours after you have been disconnected.  
What is the IP address, and MAC address of the server? What protocol is used if you know the MAC address of a host in another continent and you want to determine its processor type
Describe the purpose of each of the following smilies:
The following questions are only to be taken by professionals who wish to become a member of the Tartan Institute of Network Agents.
Explain the operation of the protocol that converts the data bits from one language (such as English bits) into the bits for another language (such as Chinese bits). Using this protocol, describe how the English bits of 0 and 1, are converted into NAW and AYE, in Scottish. At what location between Scotland and England does the conversion take place?
Estimate, to the nearest millisecond, the time taken for a data packet to travel from the second PC from the back in lab LB43 (Craiglockhart Campus, Napier) to the central square in Milan, Italy. Where are the likely slow points, and where are errors likely to occur?

Note that there is no need to show any working out, or mathematical formulas, as a single number will do. Assume that it is 9:15pm on a rainy Tuesday in April, and that someone has just crashed a computer in a lab adjacent to LB43 (Hint: It is more than 10ms, and less than 10 seconds).
Show, using illustrations (with good use of colour), how data packet travel from a council estate in Coatbridge to Stonybridge. How does this differ from data packets which originate from Wishaw? From Coatbridge, how might the traffic vary to the Glasgow Celtic WWW and the Glasgow Rangers WWW sites? Which is likely to be greater? 
Estimate the length of the cable, to the nearest inch, that connects Merchiston to Craiglockhart and what colour is the outer coating?
[Scottish Words]
If you ever want to become a member of the Tartan Institute of Network Agents, then you should maybe brush-up on your Scottish words.
[Ronald Villiers site]
An excellent example of a 1st generation site, with some great Scottish dialogue.
[Exam answers]
Well, if you can't answer any of the above questions, you can always check here. These were provided by a group of BEng (Hons) students at Napier, who obviously were not given enough coursework (;-))

Related links


Click on me, if you just want to see the magazine

Date: 6 August 2001
Ed: Billatnapier
Source: Chapter 8, Mastering Computing
Price: £10 (UK)/ $14 (US)
This is a special edition of the Networking Agent, and presents some of the most asked questions relating to networking. More information can be found in Mastering Computing.
[Show only magazine] [Previous edition][Next edition]
Question: Okay. So if you know so much about the OSI model, explain to me the OSI model in terms that a layman would understand.
Okay. Let's say that you've just written a book, and you've got to send the book to your publisher (PUB_ADDRESS). Also you've been told that your postal service can only take 20 pages at a time. The layers would be:
APPLICATION. The process of actually sending your manuscript to the publisher.
PRESENTATION. Before you sent it you would format it into the required format for the font, layout, graphics format, and so on.
SESSION. You would contact the publisher to say that you were about to send your manuscript, and arrange for a README document which explained the format of the manuscript. You would also define the TITLE of the book, and ask who has dealing with the manuscript (PERSON).
TRANSPORT. The manuscript is too large to be sent in a single unit, thus you would thus split the manuscript into bundles of 20 pages each. Next you would write a sequence number on each of the bundles, and also put the TITLE of the book on each bundle (just in case that the publisher is receiving other manuscripts of the same format).
NETWORK. At this layer you would look up the address of the person to which the manuscript will be delivered too. Each bundle would then be put in an envelope and addressed with the destination address (PERSON, PUB_ADDR).
DATA LINK. The postal service will now pass the envelopes from sorting office to sorting office, with their own mechanism for finding the next sorting office. The actual address of the envelope is always used to determine the next sorting office.
PHYSICAL. The envelopes are passed from place to place using automobiles, planes, bicycles, and, of course, by foot.
Is this okay?

Question: I've been told that I should not use copper cables to connect networks be-tween two buildings. Why?
Networks use digital signals. These digital signals are referenced to a local ground level (which eventually connects to the earth connection). The ground level can vary between different buildings (and can be large enough to give someone an elec-trical shock). Thus the ground connection between the two buildings must be broken. If possible for safety, and for reliable digital transmission, you should use a fiber-optic connection.
Also, copper cables can carry electrical surges (such as from lightning strikes), and airborne electrical noise. Electrical surges can cause great damage, and noise can cause the network performance to degrade (as it can cause bit errors).
If possible use fiber-optic cables for any long run of networking media. They tend to produce fewer problems, and allow for easy upgrades (as they have a much greater bandwidth than copper-based cables).
Question: How do you connect a fiber-optic cable to a connector?
It takes a little bit of skill, but basically it is just glued onto the end.


Question: And, how do you get an RJ-45 connec-tor onto twisted-pair cable?
You strip about 0.5 inch of the outer jacket and fan-out the wires in the correct order. Next you push them fully into the RJ-45 con-nector, and finally use the special crimping tool to clamp the cable, and make the required contacts. No soldering is involved.
Question: Why do they put cables in the ocean? Can't they just use satellites?
Satellites can transmit at hundred of megabits per second. The data which travels between continents is increasing all the time, and can be many tens of gigabits per second. Thus satellite systems would require hundreds of channels to be able to cope with the capacity. A fiber optic system has an almost infinite bandwidth, which is typically limited by the electronics that it uses, and having to boost the laser signal at regular intervals. Thus satellite communications provide a good short-term solution, but nothing can really beat fiber optic transmission. A big problem with fiber optic cables under the sea is that the seabed is forever moving, which eventually ends up break the cable. Thus cables have to be re-run at regular intervals.

Question: I use a Dial-up connection from home, and an Ethernet connection at work. Is there any difference in the way that my applications operate?

None at all (when you use TCP/IP communications). TCP/IP provides the interface between the networking tech-nology and the application program, and have been designed so that the networking type is transparent to the application pro-gram, so, for example, it doesn't matter to a WWW browser that you connect to a modem or over a LAN.

DEAR OSI WIZARD. I can never remember the order of the OSI layers. Can you help me?
Well, you should try and learn the layers for their functionality, but if you can't here's a few OSI mnemonics that you can learn:

All People in Scotland Talk about is Neeps, Dumplins and Potatoes. Can you think of any other?
Neeps - Scottish for Turnips.
Question: Can you define the functions of the seven layers
Well, here goes:
Application. Provides application programs, such as file transfer, print access and electronic mail.
Presentation. Transforms the data into a form which the session layer and the application layer expect. It can perform encryption, translating character sets (such as converting binary values into text for transmitting a binary program over a text-based system), data compression and network redirections.
Session. Setting up, maintaining and closing down of a session. It should not depend on any specific transport or network layer, and should be able to communicate as if the session was created on a stand-alone computer (that is, the network is transparent to the session layer).
Transport. Provides for reliable end-to-end error and flow control. The network layer does not validate that the data packet has been suc-cessfully received, thus it is up to the transport layer to provide for error and flow control.
Network. Defines the protocols that are re-sponsible for delivering the data to the required destination.
Data link. Provides for the access to the net-work media and thus builds on the physical layer. It takes data packets from the upper level and frames it so that it can be transmitted from one node to another.
Physical. Provides for the actual transmission of the binary digits.
Question: What do I need to create a basic network?
All you really need is two computers, two Ethernet NICs, a hub, and some patch cables. The patch cables connect the computers to the hub, and the hub creates the network. The computers can then simply make a peer-to-peer connection with each.
UNIX/Linux will allow you to access one computer from another, using TELNET, FTP, NDS, and so on, but you would have to assign each computer a unique IP address. As long as you do not connect onto the Internet, you can choose any IP address.
Microsoft Windows uses its own protocol (NetBEUI) to make a peer-to-peer connection (with file/printer sharing).
Question: I live in Edinburgh, and my friend lives in London. How long does it take for a digital pulse to travel from Edinburgh to London?

Well, there are a lot of assumptions to be made. First we'll assume that there are no intermediate devices in the cable that connects Edinburgh and London, and we'll assume that it is fiber-optic cable, which propagates light pulses at one-third the speed of light (108 m/s). Thus for a distance of 500 miles (804.65km,) the time will be:
Thus it would only take 8 thousands of a second. Pretty fast, compared with 1 hour by flight, and 6 hours by car.

What are the main rules that I should use when I install network cables?

Well, the initial installation is important as well installed cable will reduce the likelihood of problems in the future. Cabling problems tend to be one of the top causes of network prob-lems. The rules can be summarized as:
Untwisting cables. The maximum amount of untwisted in a Cat-5 cable is ½in; this is to maximize the cancellation effect.
Cable bend. The maximum cable bend is 90°.
Staples should never be used as these pierce the outer jacket of the cable. Attach cable ties to cables going on the same path, but never secure them too tightly. If possible secure the cable with cable ties, cable support bars, wire management pan-els and releasable Velcro straps.
Try and minimize outer cable twists and stretching the cable, and never allow the cable to become kinked, as this changes the characteristic impedance of the cable. The cables within can untwist when stretched.
Leave enough cable at each end so that it can be properly terminated. It is less expensive to add an extra few meters onto the length at either end, than it is to have to re-run the whole cable. Typically, the cable run will have an extra few meters hidden below in the floor, or above in the ceiling, in order to compensate for extra lengths.


DEAR OSI WIZARD. Question: Why is the OSI model so important?
Well very few networking technologies actually fit exactly into the OSI model, but the main advantages of it are:
Modular engineering.
Interoperable technology.
Reduced complexity.
Simplified teaching and learning.
Standardized interfaces.
How do I choose the right cable?
The main factors include:
Bandwidth (such as bit rate). Fiber-optic cable has the greatest bandwidth, fol-lowed by coaxial, and then by twisted-pair.
Location (such as flammable environment, security).
Expense. Fiber-optic cable is the most expensive, and twisted-pair the least.
Long-term use (upgradeability over the years).
Question: What's the difference between a network address and a physical address? And what's an IPX address?
The MAC address identifies the physical address of the NIC, and differs from the network address (which is also known as a protocol address) which is used by the network layer. An Ethernet address takes the form of a hexadecimal number, such as:
0000.0E64.5432 or 00-00-0E-64-54-32
and the network address, for IP, takes the form of a dot address, such as:
All computers that connect onto the Internet must have a unique IP address.
IPX addresses (for Novell NetWare) use an eight-digit hexadecimal address for the network address and the node portion is the 12-digit MAC address. For example:
F5332B10 is the network address, and the rest of it is the MAC address.
Question: Everywhere I read, it says that Ethernet has so many problems, and isn't really a very good networking technique. So why is it so popular?
Local area networks have evolved over the years. At one time the big contest was between Token Ring, and Ethernet. Which was best? Well Token Ring was always better at coping with network traffic than Ethernet, especially when the network traffic was heavy. But, remember these were the days before hubs. Thus most network connections were made from computer to computer with coaxial cable. The big problem with Token Ring was when there was a bad connection or when a computer was disconnected from the network, as this brought the whole network down. Ethernet (10BASE) proved much easier to add and delete computers to and from the net-work. Thus it triumphed over Token Ring. Soon Ethernet NICs cost much less than Token Ring cards, and were available from many sources (typically, these days, Token Ring cards will cost up to over five times as much as Ethernet ones).
Ethernet has coped well with the evolving networks, and the new hubs made it even easier to connect computers to a network. It faced a big problem, though, when the number of users of a network increased by a large factor. Its answer to this was 100BASE, which ramped up the bit rate by a factor of ten. This worked well, but it suf-fered when handling traffic over wide areas. Ethernet had a final trump card: 1000BASE, which gives a bit rate of 1Gbps.
Thus, whatever we throw at Ethernet, it fights back by either ramping up the bit rate (from 10Mbps to 100Mbps to 1Gbps) or it allows multiple simultaneous network connections (through Ethernet switches). So, don't dismiss the King, he's going to be around for a while yet.
Question: Everyone keeps telling me how good network switches are. So what's so good about them?
Network switches, unlike network hubs, allow more than one port to talk to another port, at a time. Thus the bandwidth can be a multiple of the base bandwidth. For example a four-port switch could have a total network maxi-mum throughput of 400Mbps (for a base transfer rate of 100Mbps). The other example of network switches is that it is possible to configure them to set up virtual LANs (vLAN), which enhances security and also allows networks to be defined by software rather than physical location.

Question: If we're transmitting voice or audio or the fiber optic cable, are we transmitting the analogue version of the signal, or a digital one?

Normally we transmit the digital version, which means that we send digital pulses of light over the fiber. It is possible to transmit different intensities of light to represent the amplitude of the signal, but this is susceptible to reduction in amplitude of the light rays.

Question: What? The light pulses bounce along the cable?

Yes. They do. The light pulses do not travel straight through the fiber, they bounce along it. Just imagine that you're underwater. When you look at the water, at a certain angle, you can actually see a mirror image of the sky. It's the same concept, really.