Basic Concepts

Types of signals - Analog and Digital

An analog signal is based on power that varies continuously over time. In the case of electrical signals, the power is usually modulated by adjusting the voltage of the power source.

Slight changes in the signal determine slight changes in the data. Consider, for example, slight changes in a sound that you hear. If the amplitude of the signal increases slightly, you hear a slightly louder sound. If the frequency of the signal increases slightly, you hear a slightly higher pitch.

The amplitude is the size of the signal from peak to peak.

The wavelength is the length of time over which a complete wave cycle is transmitted. It is measured in seconds.

The frequency is how often a wave pattern repeats. It is measured in cycles per second or Hertz and it is the reciprocal of the wavelength.

The phase is determined by when the wave passes through the origin. One wave may be shifted with respect to another because it starts at a slightly different time. Phase is measured in degrees or radians. Consider, for example, the difference between walking and jumping. When you walk, one leg is out of phase with respect to the other by 180 degrees or pi radians. When you jump, both legs are in phase.

Analog waveforms can be added together to make complex signals. Moreover, they can be shifted away from the origin by adding a constant voltage. A complex analog signal might look as follows:

A digital signal is based on voltages that vary in fixed increments or steps. A transition to a new state represents one bit of data.

Although we often think that digital signals have only two states, that is, off and on, they can have many. A digital signal might look as follows:

Because the digital computer stores data as discrete bits, a digital signal is convenient for data communication. Moreover, digital communication is less susceptible to errors because electromagnetic noise can be detected and corrected as it causes the signal to vary from its discrete states.

Types of Transmission - Serial and Parallel

Data in a digital computer is usually represented in 8, 16, 32 or 64 bit increments. For example, the number 733 can be represented in 16 data bits as 0000001011011101. When serial transmission is used, each data bit is sent individually over a single channel, one after the other. When parallel transmission is used, several data bits are sent at once. This means that if, for example, 8 data bits are sent at one time, 8 channels are needed.

Serial Transmission

Parallel Transmission

Data Transfer Modes - Asynchronous and Synchronous Transmission

For the receiver to interpret a transmission, it must know where the transmission starts and stops. In other words, the receiver must be synchronized with the sender.

If the sender is sending small blocks of data usually with intervening pauses, then asynchronous transmission is often used. Here, each eight bits of data have an extra leading bit and an extra trailing bit. The leading bit is called the start bit. The trailing bit is called the stop bit. The start bit readies the receiver to read the data bits. The stop bit simply resets the receiver to look for the next start bit. Because of the start and stop bits, the efficiency of asynchronous transmission can not exceed 80 percent. Asynchronous transmission is often used to transmit keyboard characters from a terminal to a mainframe computer.

Asynchronous Serial Transmission

If the sender is sending larger blocks of data, with or without intervening pauses, then synchronous transmission is usually used. Here, the blocks of data can either be simply a collection of bits or octets (8 bits). The blocks of data typically start with some type of SYNC or synchronization field that is used to notify the receiver that data follows. Because the blocks of data can be very long and the synchronization field is quite short, this transmission technique is more efficient than asynchronous transmission. Transmission efficiency for Ethernet (IEEE 802.3) can be greater than 99 percent.

The Universal Serial Bus (USB) on a workstation uses this form of synchronous transmission.

Synchronous Serial Transmission (IEEE 802.3)

Another form of synchronous transmission is achieved using a separate control signal in a separate channel to synchronize the sender and receiver. The parallel printer port on a workstation uses this method.

Transmission Media and Methods

Digital signals can be transmitted using copper wire, optical fiber and some other media. Digital transmission is possible in only the base band, that is, a band of frequencies that start at zero Hertz. Most media, however, are most efficient if they are used to transmit an analog signal. In this case, digital data must be transformed into an analog signal using a technique known as modulation. Analog transmission is usually referred to as being narrowband, where there is a narrow spread of frequencies, or broadband, where there is a wide spread of frequences.

A modem is a common device that performs two functions: modulation and demodulation.

There are three common types of modulation.

Duplexity - Simplex, Half-duplex, Full-duplex

There are three types of communication:

The duplexity may be determined by one of two factors: a) the physical communication channel may be only be able to support certain types; or b) the protocol, i.e., the rules, used for communication among computers may limit the type of communication although the physical communication channel may not.

Topology

Topology refers to the manner in which a network is physically arranged.

There are two types of devices in all types of telecommunication. Data Terminal Equipment (DTE) refers to the sender and the receiver of data, that is any terminating device on the telecommunications channel. Data Communication Equipment (DCE) refers to any telecommunication equipment that is connected to a DTE to transmit data.

Although you may think that a given device is a DTE, it may have an interface that is configured to be a DCE or, similarly, a DCE may have an interface that is configured to be a DTE. The interface configuration matters because the wiring between the devices is usually different for different configurations. The circuit that transmits data at one of the devices must be the circuit that receives data at the other. If two DTE's are connected (or two DCE's are connected), regardless of their functionality, a cross-over connection must be made between them.

The figure below shows the circuit wiring of a of cross-over cable for connecting two PC workstations (DTE's) by their serial communication ports. A cross-over cable connecting two workstations together is also called a null-modem cable.

Circuit Diagram for a Null Modem Cable with DB-9 Connectors

The following table shows various DTE's paired with their respective DCE's.

DTE	Cable	DCE
workstation com port	serial	modem
workstation Ethernet card	twisted pair	Ethernet hub, switch, router or bridge
telephone, modem, wall jack	quad	—

Network Components

Network Components

Network Classifications

Networks are classified by the geographical area they serve.

Local Area Network (LAN) -

• usually within a building or a cluster of buildings;
• usually private, that is, without need of a telecommunication provider;
• usually uses copper wiring as a medium;
• usually uses Ethernet (IEEE 802.3) or token ring (IEEE 802.5) protocols

LAN's are characterized by the following topologies.

Bus Topology Ring topology   Star Topology Tree Topology To all external appearances, an Ethernet 10BASE-T LAN, which uses a star topology, and token ring LAN appear to be the same. Both usually have a DCE that is connected by a cable to each workstation. The cable contains two twisted pairs of wires, either unshielded or shielded with tinfoil. The circuits in the DCE, however, are wired in an entirely different way. Token Ring (IEEE 802.5) Topology

Municipal Area Network (MAN) -

• usually within a city or municipal area;
• not yet available in most places;
• usually requires a telecommunication provider;
• usually uses fiber optics as a medium;
• usually uses high speed synchronous High Level Data Link Control (HDLC), Synchronous Optical Network (SONET), and/or Asynchronous Transfer Mode (ATM) protocols

Wide Area Network (WAN) -

• usually beyond a single municipal area;
• usually requires a telecommunication provider;
• usually a corporate initiative;
• usually uses fiber optics as a medium;
• usually uses high speed HDLC, SONET and ATM protocols

Backbone Network -

• consists of the 13 root servers of the Internet;
• uses fiber optics as a medium;
• provides a connection among all Internet subnets;
• provides address and name resolution for servers and clients
• uses high speed HDLC, SONET and ATM protocols

Application Architectures

Applications that require a network have several possible client-server based architectures. They can be classified according to the degree of functionality assumed by the workstation with respect to one or more servers.

The host-based architecture places most of the functionality on the server. The client is a terminal or workstation that behaves like a terminal.

Host-based Architecture

The client-based architecture places most of the functionality on the client. The server stores the data but the client performs all of the application, data access and presentation functions.

Client-based Architecture

The client-server architecture more evenly divides responsibilities between the client and the server. In this case, the server stores the data and performs the various data access functions whereas the client performs the application and presentation functions.

Client-server Architecture

The web architecture places all but the presentation function on the server. Consequently, the server stores the data, provides data access services and supports the application logic. The web architecture is being used by the SCHIP demonstration system.

Web Architecture

The following table makes a comparison between the different application architectures.

	Host-based	Client-based	Client-server	Web
Cost of Infrastructure	high	medium	low	low
Cost of Development	low	medium	high	medium
Scalability	low	high	medium	medium

Network Architectures

The term network architecture refers to the protocols or sets of rules followed by interconnected telecommunication equipment. Protocols define the

• syntax - data format, coding, and signal levels,
• semantics - control information for coordination and error detection and correction, and
• timing - speed matching and sequencing

The protocols used by the Internet are are designed to work with the networks composed of many computers. Some of the computers are interior nodes (also referred to as Internet Message Processors or IMP's) that switch data packets through the system to reach hosts (or clients) that are at the boundary of the network.

Schematic of a Packet Switched Network

There are two common models used to describe network architecture. Both models demonstrate the concept of separation of concerns by isolating specific functionality to specific layers. This allows for easier conceptualization and specification. Many of the software implementations of these models use the same approach in order to simplify programming and testing.

• The International Standards Organization (ISO) has developed an idealized model named the Open Systems Interconnection (OSI) Reference model that consists of seven layers. Each layer has a set of functions and provides a set of services. Although systems have been built in accordance with the model, these systems are not in as common use as is the TCP/IP model.

• The Internet model corresponds roughly to the OSI model but lacks some of the layers specified in the OSI model. The same functions and services are supplied but they are grouped in a different way. The Internet model is used to describe the network architecture of the Internet. The following table compares the OSI Reference model to the Internet model.

OSI Reference Model Compared to the Internet Model

Conceptually, the data is exchanged as bits at the physical layer. Then each layer operates on the data in succession. Each layer provides its own control information using the concept of encapsulation. In the following figure, the ISO Reference model is shown illustrating the exchange of data between two stations. At the sender, when data is sent by one application to another, a header is added to the data unit as the unit is passed down through the protocol stack. This header contains instructions to be given to the corresponding layer at the receiver. The receiving application is given only the data that is sent by the sending application; that is, the control information is extracted by each respective protocol before the data unit is passed upward to the layer above.

ISO Reference Model Using Encapsulation

There are a large number of protocol standards. A protocol usually applies to only one layer or to one component within a layer. The following is a list of some protocols showing where they are positioned within the OSI Reference model.

OSI Layer	Examples of Protocol Standards
7. Application Layer	FTP (file) HTTP, HTML (web) MPEG, H.323 (multi-media) IMAP, POP, X.400 (e-mail) LDAP, X.500 (directory)
6. Presentation Layer	BER (Basic Encoding Rules) ANSI X3.4 (ASCII encoding) UTF-8 (Universal Character Set encoding) ASN.1 (Abstract Syntax Notation)
5. Session Layer	ISO 8826, 8827
4. Transport Layer	TCP (Transmission Control Protocol) SPX (Novell Sequenced Packet Exchange)
3. Network Layer	IP (Internet Protocol) IPX (Novell Internetwork Packet Exchange)
2. Data Link Layer	Ethernet (IEEE 802.3) Token ring (IEEE 802.5) PPP (Point-To-Point protocol, RFC 1661)
1. Physical Layer	Serial (RS232C) V.92 (56 kbps modem) TSB-36 (unshielded twisted pair standards)