http:\\www.wholewiredworld.org\networking.htm\

Basic Networking

Basic Networking

Networking is by far one of the most quickly expanding fields in the computer industry. With the surge in popularity of the Internet in the mid 1990’s, a great deal of interest in networking has arisen. Whenever two or more workstations are connected together, a Local Area Network (LAN) is created. Nearly every major company has, at a minimum, a LAN installed.

There are some basic networking concepts that you will run into time and time again as a technician. Whether you have an interest in networking or not, it will be in your best interest to be familiar with these basic concepts.

After you have a firm grasp of some basic networking concepts, it is also important to have a general understanding of how to configure a Network Interface Card (NIC). Configuration of a NIC is extraordinarily simple in most cases, and will be useful to you during your career.

Basic Networking Concepts

As a base for your networking abilities, you need to understand some basic networking concepts. Everything in the world of networking revolves around these concepts, and a firm grasp of them is extraordinarily important. These basic concepts include:

Cabling

Network Interface Cards

Full Duplex Network Access

Protocols

Cabling

Obviously, in order to create a network, you have to somehow physically connect the devices that will be on the network. This is accomplished using cables. There are many different types of cables, each having it’s own advantages and disadvantages. In the next few sections, we discuss three types of cabling: Twisted Pair, Coaxial, and Fiber Optic.

Twisted Pair

Twisted pair is by far the most common type of network cable, primarily because of its low cost. Physically, twisted pair consists of pairs of wires, usually four. Each of the wires in a pair is wrapped around the other to help avoid interference. Twisted pair requires that each workstation be attached to a hub (see Figure 1-1). A hub receives data from one of its ports, and then transmits it to all of its ports. Twisted pair cable is considerably less expensive than the other types of cabling. In larger network installations, the lower price of cable offsets the cost of a hub and makes twisted pair the most cost-effective networking solution.

Figure 1-1: A hub is a device to which each device on a network connects

Twisted pair

Twisted pair can be referred to by many different names. Some of these names include unshielded twisted pair, UTP, shielded twisted pair, 10BaseT, or 100BastT; UTP is the most popular. As shown in Figure 7-2, twisted pair is configured in a star topology, in which each device is connected to a central device, usually a hub. In the event of a cable being cut or broken, the device that is connected to that cable will no longer be able to communicate on the network, but will not effect any other devices on the network.

Figure 1-2: A twisted pair network is usually configured in a star topology

Coaxial

Coaxial, while not as common as twisted pair, is also a popular type of network cable. Physically, coaxial cable consists of a central wire that is surrounded by a screen of fine wires. Coaxial cable is most common in smaller networks, where it is cheaper to purchase the coaxial cable instead of purchasing both the UTP cabling and hubs. Like twisted pair, coaxial can also be referred to by many names. Some of these names include BNC, ThinNet, and 10Base2; ThinNet is the most popular. Each device must be connected to a T-connector, shown in Figure 7-3. Each T-connector is connected to the next with a coaxial cable. After all of the devices are connected, the ends of the cable must then be terminated with a 50W terminator. As shown in Figure 7-4, a coaxial network is configured in a bus topology. In the event of the cable being cut or broken, the network will cease to operate.

Figure 1-3: A T-connector is required between each device on a network utilizing coaxial cabling and the coaxial cable

Figure 1-4: A coaxial network is usually configured in a bus topology with each device connected to a main cable or bus

Fiber Optic

Fiber optic is by far the least common of the three types of cabling. Fiber optic cabling is usually found where long cable lengths are required, where extremely high speed is desired (at extremely high cost), or where there is high EMI radiation or other environmental difficulties. Traditional copper-based cables such as twisted pair or coaxial cables are very susceptible to radiation and other environmental difficulties. Fiber optic cable is much less susceptible to these environmental difficulties because it uses light signals rather than electrical signals. Light signals offer much more protection from environmental interferences than electrical signals. Fiber optic is used when long lengths of cable are required because it is able to sustain longer distances without environmental interruption. Fiber optic is usually referred to simply as fiber. Each segment of fiber optic cable must be connected at each end with a special fiber optic connector. As shown in Figure 1-5, a fiber optic network is configured in a ring topology. In the event that the fiber optic cable is cut or broken, the network will cease to operate. (Most fiber optic installations actually include two rings in order to provide redundancy and fault tolerance.) Table 1-1 lists and compares each of the cable types discussed in this section.

Figure 1-5: A fiber optic network is usually configured in a ring topology where a token is passed around a ring

Cable Type	Topology	Maximum Distance	Transmission Speed	Other Required Devices
Twisted Pair	Star	100 Meters	10 or 100 Megabits per second	Each networked device must connect to a port on a hub.
Coaxial	Bus	180 Meters	10 Megabits per second	Each networked device must be connected with a T-Connector and terminators must be connected to each end.
Fiber Optic	Ring	Theoretically no limit	Theoretically no limit (100 Megabits per second most common)	Each end of a cable segment must have a special fiber optic transmitter connected.

Table 1: Comparison of Important Characteristics of Cable Types

Network Interface Cards
Network interface cards (NICs) are the devices that, when installed in a PC, connect the PC to the network cable. NICs must be matched to the type of network and the type of network cable that you are using. It is important to choose a NIC that corresponds to your network. Recently, most NIC manufacturers have introduced models that can connect to either twisted pair or coaxial networks.

Full Duplex
Networking is described as full duplex when it is capable of transmitting in both directions at the same time. In simpler terms, a full duplex network medium has the capability of sending and receiving simultaneously. The most common application of true full duplex networking is in fiber optic cabling.

Network Access
Network access (or access methods as it is more commonly referred to) is the method by which devices communicate on the network. Without some type of control on the network, any device could transmit at any time, quite possibly at the same time as another device. As we humans have a problem when two people attempt to communicate simultaneously, devices on the network have a similar problem. Network access provides a standard that all devices that wish to communicate on the network must abide by in order to solve this problem. Common types of network access are:

Carrier Sense Multiple Access/Collision Detection
Carrier Sense Multiple Access/Collision Detection (CSMA/CD) is somewhat similar to how we humans communicate. With CSMA/CD, a device transmits data onto the network. The device then detects if any other devices have transmitted onto the network at the same time. If it detects that another device has transmitted data onto the network, the device then waits an unspecified random amount of time and retransmits its data. CSMA/CD is most commonly found on Ethernet networks, which most commonly use twisted pair or coaxial cable as the physical cable type. CSMA/CD is one of the faster access methods available, but performance can become an issue on busier networks.

Token Passing
Token passing is a bit more organized than CSMA/CD. Using token passing, a token is passed from device to device around a virtual (and frequently physical) ring. Whenever a device receives the token, it is then allowed to transmit onto the network. Token passing is most commonly found on fiber optic and token ring (which we will not discuss because it is becoming less commonly used) networks. Token passing, although more organized than other methods, is a bit slower because it requires that devices wait until they hold the token prior to transmitting onto the network. In the event that the token were to be lost, the network could be rendered useless until a new token is created, which could take from a few seconds to a few hours. The use of token passing is advantageous where total organization is required, such as in fiber optic backbones.

Protocols
Data cannot simply be transmitted onto the network. Protocols establish standards for such transmission of data. Both the sending and receiving device must be capable of communicating using the same protocols. Some common protocols are:

TCP/IP
TCP/IP, Transmission Control Protocol/Internet Protocol, is the most common protocol used today; it’s the protocol upon which the Internet is built. TCP/IP was originally designed in the 1970's to be used by the Defense Advanced Research Projects Agency (DARPA) and the Department of Defense (DOD) to connect systems across the country. A requirement of this design was for the ability to cope with bad network conditions. An advantage of TCP/IP is that it is routable or can be passed beyond a router. A router is a network device that connects two or more networks together. The router that all traffic it allows to cross is destined for the network on the other side of the network prior to allowing it to cross. TCP/IP’s largest disadvantage is that it requires quite a bit of configuration prior to use.

IPX/SPX
IPX/SPX, Internetwork Packet Exchange/Sequenced Packet Exchange, is the protocol most commonly used with Novell Netware. IPX/SPX is a very fast and highly established protocol, but is not used on the Internet. Novell developed IPX/SPX for use in Netware. Like TCP/IP, IPX/SPX is routable and requires some configuration, though nowhere near as much as TCP/IP.

NetBEUI
NetBEUI, NetBios Extended User Interface, is a transport protocol that is commonly found in smaller networks. NetBEUI is an extremely quick protocol with very little overhead that was first implemented with LAN Manager products. NetBEUI is not routable and requires little configuration, if any at all.

Ways to Network a PC So far, all of our discussion has involved using a NIC to connect two or more computers together. There are a few other ways to network workstations.

Ways to Network a PC
So far, all of our discussion has involved using a NIC to connect two or more computers together. There are a few other ways to network workstations.

Dial-Up Networking is when a modem is used to connect two or more workstations together. The majority of people who connect to the Internet from home use Dial-Up Networking on a daily basis. It is considerably slower than using NICs, but can accomplish the same tasks.

A direct cable connection is when two computers are networked using either a serial or parallel cable. Direct cable connections are considerably faster than modem connections, but are still slower than networking using a NIC. Direct cable connections are limited to networking a maximum of two computers and are limited by physical cable restrictions.

Configuring Network Interface Cards
Network Interface Cards (NICs) are fairly simple to configure. Most NICs require an IRQ and an IO memory address. Once the resources that the NIC uses have been configured on the NIC (most newer NICs use a software utility rather than the more traditional jumpers or dip switches; consult your NIC’s documentation for exact specifications), the Network Operating System must be configured with the same settings. Configuration will be a less difficult process if the same type of NIC is used in each workstation on your network.

Problems on the Network
As the saying goes, all good things must come to an end. Someday your beautiful network will not continue to operate exactly as it used to. This could be the result of many different things. Some common causes of network problems include:

* Physically damaged cable
* Damaged Network Interface Card
* Excessive traffic on the network
* Incorrectly operating hub
* Missing or incorrectly operating terminator
* Magnetic fields
* Incorrectly configured network devices

Reduced Bandwidth
Reduced bandwidth occurs when users’ data transmissions across the network begin to take longer, and begin to be timed-out by the applications requesting the transmission. Reduced bandwidth can be caused any of the common network problems mentioned in the preceding section. The most likely cause of reduced bandwidth is excessive traffic on the network. Excessive traffic usually is not a result of an equipment failure, but instead a result of equipment success. The only solution to excessive traffic is to modify the network configuration in order to allow more traffic, or to separate high-traffic users from others. The most common method of modifying the network configuration to allow more traffic is to upgrade a 10 Megabits/Second LAN to 100 Megabits/Second by replacing hubs, NICs, and possibly the cabling.