Section 10 - Communications and Networking Basics

Networking terminology:

Peer-to-peer network: This type of network includes two or more PCs connected to share data files, a printer, or other resources.
Local area network (LAN): A small business or corporate department may install a LAN interconnecting from two to hundreds of PCs, using permanently installed cabling or perhaps a wireless technology.
Wide area network (WAN): A corporation may maintain a WAN using dial-up, leased, or other dedicated communication means.
Nodes: Computers on a network are sometimes called nodes. Computers and devices that allocate resources for a network are called servers.
Topology: The geometric arrangement of any network is its topology. The most common topologies are the bus, star, and ring topologies.
Protocol: To operate efficiently, any communications based system must have established rules to govern its operation. This set of rules is its protocol. Popular protocols for LANs are Ethernet, token ring, and FDDI.
Data packets: Data, messages, and tokens that are transmitted on any network must conform to the size and format prescribed under the network's protocol. Data packets can also vary between network operating systems (NOS) on the same protocol.
Architecture: Any network can be classified as either a peer-to-peer or client/server architecture.
Media: Nodes on a network are connected by twisted-pair copper wire, coaxial cable, fiber-optic cable, or wireless radio wave connections.

All phases of network operations are defined in the layers of the Open Systems Interconnection (OSI) model developed by the International Standards Organization (ISO). The ISO OSI model has seven layers:

The physical layer is the bottom and most technical layer of the OSI model. This includes cables, their media, length, connectors, and so on. The physical layer also defines the characteristics of the electrical signals used to transmit data around the network without assigning any particular meaning to the signals.
The data link layer assigns meaning to the electrical signals tranmitted around a network. This layer determines the size and format of the data sent between the PCs, printers, and other devices (all of which are called nodes) on a network. It also defines the error detection and correction schemes.
The network layer provides definitions for the interconnection of two possibly dissimilar networks so that data from one can flow to the other.
The transport layer breaks data into easily handled message sizes and also provides the identification and addressing that allows a message to find its way from one node to another.
The session layer carries out the communications between nodes exchanging data. A session must be established before an exchange of data can take place.
The presentation layer is primarily responsible for the coding and decoding of either compressed or encrypted data sent over the network.
The application layer provides the linkage for application programs to interact with the network.

A LAN can be connected to another LAN to form a bigger LAN. LANs can be interconnected indefinetly, but as long as they are contained within a single physical area, they remain a LAN. When LANs, or nodes, are connected over longer distances, such as miles, states, or continents, they form a wide area network (WAN). WANs are usually constructed using ISDN, leased lines, or other high-speed connections.

The three basic types of network topologies are the star, ring, and bus topologies:

The star topology is common to ARCNet (Attached Resource Computer Network) networks. Every node in a star network is directly connected to a central server.
The ring topology, which is known more commonly under its commercial name of IBM's Token Ring, creates a network structure in which all the nodes are connected one to another to form a loop that starts and ends at one of the nodes, possibly the server.
The bus topology, or Ethernet networking, is the most common network topology in use. All nodes are attached to a backbone (the main Ethernet line).

Fundamental Features

     Networks have essentially five parts: nodes, NICs, cables, connectors, and a network operating system (NOS).

The Network Interface Card (NIC)

     The NIC is a physical and logical link for a PC to a network. The NIC is installed inside the computer into an unused expansion slot. The NIC is a translator that works between the network and the PC. Networks transmit data in a serial data format (one bit at a time), and the data bus of the PC moves data in a parallel format (eight bits at a time). The NIC converts the signal from serial to parallel or parallel to serial, depending on the direction. The NIC also formats the data as required by the network architecture.

Cabling

     Three basic media are used for network cables: coaxial cable, fiber-optic cable, and cables consisting of twister pairs of copper wire. You need to know each type of cable and the network architecture it is most commonly used with:

Coaxial cable come in two types for networks:
- Thick coaxial cable, which is usually yellow in color and used in what are called thicknets, have two conductors. It is made of a solid center wire and a braided metal shield. It can be used in 500 meter lengths.
- Thin coaxial cable, used in thinnets, have the same basic construction of thick coaxial except it is made of lighter and less expensive materials. This cable is also referred to as RG-58 A/U. The A/U indicates that the core of the wire has a stranded cord (strands of wire wound together).
Fiber-optic cable, or "fiber" has several advantages. It carries a huge amount of data very fast, is small in size, resists corrosion, and is very reliable. On the other hand, it is hard to splice, difficult to install, and it and the devices that connect to it cost more than other types of networking cable.
Twisted pair cable is available in two types: unshielded and shielded:
- Unshielded twisted pair (UTP) is the most popular type of network cabling in use. UTP has four pairs of wires inside a plastic sheath. Because each of the copper wires is only covered with a thin plastic sheath, each pair of wires is twisted a different number of twists per inch to eliminate interference from the other pair of wire or other electrical devices near the cable.
- Shielded twisted pair (STP) is common in higher speed network architectures. STP is very much like UTP with the added bonus of a metallic shield wrapped around the wire bundle to eliminate interference from outside devices, such as electrical motors, fluorescent light fixtures, etc.

Twisted pair cabling is defined in categories or grades that provide an increasing amount of protection for outside electrical noise and interference. Networks should be created with cat 3 cable or better, with cat 5 preferred.

Connectors

Different types of connectors are used with different types of cable media. While serveral special connectors are used to connect network cables to NICs and other network hardware, the two most commonly used are the following:

The BNC (Bayonet Naur Connector) is used to connect thicknet network cables to the NIC.
The RJ-45 connector looks much like its cousin, the RJ-11 connector, which is used on telephone lines. The RJ-45 connects eight wires, whereas the RJ-11 connects four or six.

Network Architectures

NICs come in different varieties to match the architecture of the network to which they connect: Ethernet, token ring, or FDDI:

Ethernet, or bus topology, uses a full range of network media (copper or fiber) and operates at either 10 or 100 Mbps. The 100Mbps is called Fast Ethernet. An Ethernet network can support about 500 nodes.
Token ring, or ring topology, also uses copper and fiber-optic cabling, operates at 4 to 16 Mbps, and supports about 260 nodes. A token ring network operates very reliably but can be hard to troubleshoot.
FDDI stands for Fiber Distributed Data Interface. A FDDI NIC contains a laser transceiver that converts its digital data into light to be transmitted on a fiber-optic network or back to digital signal form incoming light impluses for use by the PC. FDDI is a standard of the American National Standards Institure (ANSI) and the International Standards Organization (OSI) for data networks using ring topology and data speeds of 100 Mbps.

The component on the NIC that handles the data transmission duties is the transceiver. Some NICs have more than one transceiver, common on Ethernet cards, with each attached to a different connector type on the card.

Ethernet Networks

     Ethernets are easily the most common network type in use. Although it is generally based on the bus topology, it can also be installed as a star. An Ethernet network can also be clustered around one or more hubs in a star design. A hub is a device that connects many devices to a server through a single line.

     Ethernet networks are passive, which means that the nodes do all the work. To avoid data collision, an Ethernet node checks the line to see whether is is in use before sending a message. The protocol used for this purpose by Ethernet devices is carrier sense multiple access with collision detection, or CSMA/CD.

Ethernet Cabling

     The IEEE 802.3 standard defines the various versions of Ethernet and its cabling. The cable systems used in Ethernet are as follows:

10Base2, also called thin Ethernet, thinwire, or thinnet, uses light coaxial cable to create a 10 Mbps network with cable segments that cannot be more than 185 meters in length. 10Base2 cables connect with BNC connectors. The NIC uses a T-connector where the cables from two adjacent nodes connect. Ethernet requires that any unused connection must have a 50-ohm terminator.
10Base5, also called thick Ethernet, thickwire, or thicknet, is the original cabling standard for Ethernet using coaxial cables. The names comes from its 10 Mbps speed and its maximum cable length of 500 meters between nodes. 10Base5 connects with BNC or AUI (attachment unit interface) connectors.
10BaseT, where the T represents "twisted", uses UTP cable to create a 10 Mbps network with cable segments limited to 100 meters. Because it is the least expensive and easiest to work with, 10BaseT Ethernet networks are quite common.
100BaseT is also called Fast Ethernet and uses STP cable to create a 100 Mbps network.
10BaseF uses fiber-optic cable to create a 10 Mbps network. Fiber-optic based networks use special connectors, hubs, and NICs to create a network. Fiber-optic cable is commonly used to interconnect between buildings on a network, with copper used to connect the nodes to the network.

Token Ring Networks

The Ethernet is a collaborative development, but Token Ring is an IBM development that is now covered by IEEE 802.5. Its basic premise is that the network's primary cable (backbone) is arranged in a ring or loop. The control mechanism used in a ring structure is called token passing.

The token, a special packet of data, is passed from PC to PC, node to node, on the network. Any PC that wants to send a message must wait until it has the token, at which time it can send its data. Nodes connect to the loop through devices called multistation access units (MAUs). MAUs are most commonly found on large corporate mainframe sites. Token ring networks are actually configured more like a star with the MAU used as the clustering device. Becuase of its physical star configuration, any single station that breaks down has little effect on the ring, as long as the MAU can omit it from the loop.

FDDI Networks

The FDDI (Fiber Distributed Data Interface) uses fiber-optic cabling to create a network with data transfer speeds of 100 Mbps. A FDDI can support 1,000 nodes and has the flexibility of having workstation nodes as far apart as 2 kilometers.

Networking Hardware

A repeater amplifies the signal on a cable. Two types of repeaters are used:
- An amplifier repeater picks up the incoming signal and, without making any modifications to it at all, amplifies it and sends it on - good, bad, or noisy.
- A signal-regenerating repeater sends out a cleaned up signal by making a fresh and exact copy of just the good part of the incoming signal.
A hub is used as a clustering or connecting point for nodes on an Ethernet network. In many ways, the MAU of the token ring network works much like a hub.
A bridge is used to interconnect two or more LANs or LAN segments.
A router is an intelligent device that examines the source and destination addresses of network traffic and decides where the message needs to be routed. The job of the router is to determine the best possible way to send the message to its destination. Not all networking protocols support routing.
A gateway connects networks that use different protocols. A gateway is used to interconnect an Ethernet network to a token ring network. The gateway must translate and reformat the data packets from the sending protocol to that of the receiving protocol.

Protocols

Protocols
Protocol/Layer	Acronym	Description
Point-to-point Protocol (physical/data link layer)	PPP	Used to connect and manage network communications over a modem.
Transmission Control Protocol/Internet Protocol (network/transport layer)	TCP/IP	The backbone protocol of the Internet; very popular for intranets, internal networks using the Internet, and Web work-alikes.
Internetwork Packet Exchange/Sequenced Packet Exchange (network/transport layer)	IPX/SPX	The standard protocol of the Novell network operating systems.
NetBIOS Extended User Interface (network/transport layer)	NetBEUI	A Microsoft protocol that is used only by Windows OS for LANs with no external connections; does not support routing (addressing through a router to other networks)
File Transfer Protocol (session/presentation layer)	FTP	Used to send and receive files in a client/server mode to or from a remote host.
Hypertext Transfer Protocol (session/presentation layer)	HTTP	Used to send WWW documents across a network.
Network File Services (session/presentation layer)	NFS	Allows the network node to access network drives as if they were local drives, files, and data; also performs the file access and data retrieval tasks requested of the network.
Simple Mail Transfer Protocol (session/presentation layer)	SMTP	Used to send electronic mail (e-mail) across a network.
Telnet (session/presentation layer)	-	Used to connect to a remote host and emulate a terminal that the remote server recognizes.

Servicing a Networked PC

If you determine that the PC is part of a network, follow these steps to perform the repair:

Check to see if the PC is logged on to the network.
If you are working on the hard disk, make a backup of all files.
Log off the PC as necessary.
Disconnect the network cable from the NIC card and proceed with the repair.
After the repair is complete, reconnect the network cable, verify that the network files are on the disk, and restore them if needed.
Have the customer log on to the network to verify that all is well.