Networks

Introduction

What is a network?

A network is the linking together of data processing nodes such as computers (clients, servers and peers), peripherals and telephones for communication purposes. The nodes are linked together by wires or electromagnetic radiation.

Why do we need a network?

Computers, people and peripherals to name a few sometimes need to communicate with each other because it helps:

  1. to share resources (e.g. the records held in a database or there may be a handful of printers in the area).
  2. to eliminate duplicate functions (i.e. knowing someone or something else is doing the same thing can save time and money by not repeating the same process);
  3. to increase storage capacity of a workstation connected to the server on a network (e.g. workstations are deliberately designed to have limited hard disk space thereby forcing users to access a server for the required hard disk space. Or there are situations where you don't have hard disk space on your independent computer. Having access to a server for extra storage can be a godsend).
  4. to share portions of the complex, time-consuming computer tasks with different computers. In essence, you can turn a workstation into a supercomputer simply by sending some of your work to be automatically processed by any number of workstations in the area.
  5. to increase security of information (up to a certain point so long as other people don't know the password, you can encrypt the information, and people cannot tap into parts of the network including the server with the right tools to obtain the information) by decentralising the information and storing it in different parts of the network, or by centralising the information onto a server. In the latter situation, the perceived security is mainly for the individual working on a workstation because he/she thinks the information can't be stolen from his/her office should the information be stored elsewhere on a server.
  6. to understand what one device is doing before performing a function on another device (i.e. sometimes tasks need to be performed sequentially for optimal results);
  7. to help increase the accuracy and speed of information being transmitted and received (i.e. it is better than running off your feet talking to everyone on the network — let the computer and network do almost all the work for you);
  8. to ensure information gets to someone else in a reasonably secure manner (if you use encryption and you know exactly where the person is located on the network via an IP address or some other form of network identification information);
  9. simplify backup, auditing and installation/distribution of information for all or specific machines on the network; and
  10. to ensure all computers on a network behave in a consistent manner.

In the end, it is about sharing information and saving money and time for people. And we do it because we want:

  1. to have more control over many things by seeing in real-time on a 24/7 basis what everyone and everything else is doing;
  2. to make quicker and broader decisions affecting the entire operations of a business, at home, or in your personal life (so long as the decisions are good);
  3. to protect information in the event some disaster occurs by decentralising this valuable asset; and
  4. to save time and money because of the speed of communication and how many functions can be performed and resolved without duplication or wasting unnecessary resources,

which is why networks were developed.

Are there disadvantages in networks?

Well, actually there are disadvantages believe it or not! The main problems with implementing a network are as follows:

  1. Digital devices (e.g. computers), the types of networks (e.g. wireless or not, ethernet or serial etc), and even the people who use the networks can communicate differently. In other words, there is not enough consistency (actually the word is standardisation) in the language we use to communicate on the network. For example, people have different languages for communicating with others such as Spanish, German and English which we quickly become aware of through networks. And sometimes the electronic language protocols for communicating digitally on a network can also be different because of hardware manufacturers competing with each other for a share of the network market.
  2. The increasing ease in connecting one computer with another on a network means a large number of computers will become more susceptible to receiving and distributing dangerous information such as computer viruses.
  3. Computers in a network are more susceptible to being hacked with information being stolen and the culprit potentially untraceable and anonymous in the right network environment given the right black market software tools.

    NOTE: You'll need to have a very good operating system (e.g. Microsoft Windows 2000) to provide a reasonable level of security. And even then, you must be aware that no operating system is perfect.

  4. You are potentially jeopardising your privacy as more people have access to the information on your computer.
  5. There is a physical limitation in the size of the network and where you can have the server locked away. For example, information can travel only so far through a wire before the resistance of the wire and the electromagnetic noise emanating from other electronic devices in the environment washes away the signal. Hence you will need repeaters placed in various locations. Also the server is not entirely a small device. It tends to fill up an entire cupboard or small room.
  6. It can be costly to implement a reasonably secure, large and independent network (unless the network is already available such as the Internet).

Only use the network when you need to

Despite these setbacks, you need to remember the importance of keeping everything in balance (or moderation) as the old saying goes. Only use the network where it is absolutely necessary (e.g. printing a document to a network printer, sending an email to someone etc), and switch off all connections to the network as soon as you have finished with it. Ideally, this should be done automatically so you can be properly protected from digital harm from outside and to give yourself time to perform the necessary individual, creative and/or secure digital work without fear of your work being stolen or damaged.

When network access is turned on, you should always be aware of this and have given consent to do so (e.g. sending an email message or downloading a file). Otherwise it should remain off at all times. This is the second best way of making it impossible for amateur and professional hackers to break into your computer on a network or for network administrators to stop prowling around in a manner that may compromise your privacy.

If you have to keep your computer constantly accessing a network, install a quality firewall and have only the essential information you need on your network-accessable computer. This ensures people who are legitimately entitled to snoop around on the network (e.g. system administrators, managers etc) using their own special network tools are also forced to respect your privacy at all times.

Finally, you may need to ask yourself: "Do I really need to be on the network 24/7 to achieve my goal(s)?"

With so much hype of getting on the network and staying there all the time, you may be better off setting your computer to access the network every 30 seconds to obtain a piece of information each day (it should take a fraction of a second to get the piece of information you want), rather than a piece of information constantly downloading every second of the day. It will certainly reduce the information overload some people are experiencing these days with networks as well as the potential dangers of being on a network for such a long period of time with other people who you may not trust or know their real intentions.

To be absolutely safe, physically disconnect your computer from the network when you know you will not need to be on the network. This is the best form of protection you can have.

Some things to consider when building a network for someone

When a client asks you to build a network, you need to do several things:

  1. Purpose

    Always know the purpose (or scope) of the network. In other words, ask the client what they hope to achieve with the network.

  2. Mission Critical Software

    What is the mission critical software to run on the network? For example, do you want to run Microsoft Office only? Would you need to update old information and transfer it to the new network?

  3. Training

    Ask the client how much training would be required for people to use the network. This will be important when deciding on the right software for the network.

  4. Security

    What level of security do you want? The answer to this depends on how valuable the information is to the clients. But remind them that no security solution will be perfect. Given enough time, people will always find loopholes and no information can be kept secret forever. See security as a means of delaying the inevitable. The better the security, the more time you have to do what you have to do. Afterwards you must be prepared to let go of the secrets or delete what you have.

  5. Inspection

    You will need to do a physical site inspection so you can see how much it is likely to cost.

  6. Paperwork

    Get everything down in writing in what is known as a scoping document. Then write another document to let the client know roughly how much and the quantity and type of components you intend to use when building the network. This rough estimate document should have no obligations for you or the client to do anything. All it should show are three options: the cheapest, the intermediate level and the super plan version. If possible, try to show a sample machine or two running the mission critical software on a network. If the client is happy with a particular option, write an official invoice. This is the important document by which the client and you must stick to because it is a legal document.

Systems you can use

The systems available to build a network are as follows:

  1. Windows 98 can be used for servers and clients, but is only useful for low-end machines. Minimum RAM requirements to run Windows 98 is 12MB (recommended 32MB) and a minimum of a 486 computer.

    NOTE: You are recommended using the File Allocation Table 32 (or FAT32) format of Windows 95/98 to allow more files to be kept more efficiently on the hard disk.

  2. More powerful computers require Windows 2000 to create a server/client situation on a network. Minimum RAM requirements to run Windows 2000 is 32MB (recommended 64MB) and a Pentium II 266MHz machine or higher.
  3. Windows 2000 has in-built firewall, but it is not very good. Most people use Zone Alarm which is a bit better.

    NOTE: Windows 2000 comes with a new file allocation standard called the New Technology File Standard (NTFS).

  4. The best advantage of using Windows 2000 is when you need to create a secure network and allow a greater number of more powerful computers to run.
  5. Only use Windows 2000 on computers with a minimum of a Pentium II 266MHz.
  6. Use a Pentium III, IV or higher for multimedia applications (e.g. video editing, photoshop, voice recognition etc).
  7. If you choose Windows XP for networking, you have the disadvantage of having to ask Microsoft to get an activation code. Otherwise you could get into trouble. Another disadvantage is the size of the operating system — it's massive! To run it would require no less than a Pentium IV to make it bearable for users to work on a network.
  8. Not recommended for networking PCs is Windows ME (also known as Windows Millenium). This is a rushed job from Microsoft and the networking component was never completed properly when it came out in late 2000. Microsoft wanted to combine Windows NT4 with Windows 98 to create a new and more secure operating system for 2000 but realised things were getting late to complete the project at around October. So Microsoft decided to create two versions — Windows 2000, and Windows ME. It is better to go for Windows 2000 (more useful for businesses and it is much more secure to use). Windows ME are for standard consumers who think they are getting a more secure and better operating system when in truth it isn't.

General procedure in building a network

  1. Set up one workstation and connect it to the server. Use this as the test unit to help your client see how it works. When setting it up, it is recommended you put all the software on it. This is the time to see if everything works and the client is happy.
  2. If the client is happy, write down a complete list of all the components — both hardware and software — and prices to set up the network. Include the cost of labour to install the network. For a rough estimate of time, one person would normally spend at least 3 weeks to set up a small to medium-sized network system. Network specialists normally charge around A$55 per hour for this kind of work.
  3. When the client gives the go ahead to build the network, get 30 per cent of the fee paid upfront by the client. Once paid, start replicating the hardware and software configuration of the demonstration workstation for all the workstations required to access the network by the client.
  4. Use a ghosting tool to replicate software on the primary hard disk of the demo workstation to all other workstation hard disks. Ghost 2003 is a good version as it can compress hard disk information onto a CD for you to distribute to all the machines. However any other ghosting software is suitable (you will have to disconnect hard drives from workstations and connect to the ghosting machine which will require some knowledge of how internal hardware components are put together).
  5. Make sure the workstation hardware are identical or the ghosting procedure on Windows 2000-03 may not like it. Ghosting often requires going into DOS to do its ghosting. But Windows 2000 has a special version of DOS not easily recognised by the ghosting software.
  6. Unroll the Ethernet cables and tie them to the wall. Add the Ethernet socket terminals to various positions where the workstations and the server are likely to be positioned. Connect the ends of cables to the terminal sockets.

  7. Connect the server machine to the ethernet terminal socket in a secure room. For Internet access, a router and another "Internet-enabled" machine with a modem must be added to the secure room. The router has multiple Ethernet sockets designed to turn on and off the Internet access in packets to selected Ethernet cables plugged into the router. This is how users get the impression they have simultaneous Internet access on multiple workstations.

  8. To specify which terminals and hence which workstations have Internet access, the cables are plugged into sockets within the router. Each socket has a unique label to help make it easier for the network administrator to know which Ethernet terminal is "active". The router is usually locked behind a metal and glass box to prevent unauthorised users changing the connections or making new "active" terminals for people who want a free ride on the Internet and have access to the network in general.

For more specific details of how to set up a network on PCs, please read the next section.