Computer Networks

Computer Networks

    A computer network is a system for communication among two or more computers. The main advantage of a network is that data can be exchanged rapidly and easily on a long distance.

    Today’s large amount of computer networks can be classified by 3 categories. Networks can differ from each other by functional relationship, by network topology and by size of the network.

    Functional relationship type, or the architecture of the network, is a common fundamental algorithm of building the network. The simplest type of the networks is peer-to-peer, which means that different computers are directly connected to each other. Such networks are decentralized and used mainly for sharing the information between users. LAN and P2P networks are based on this architecture.

    Client-server networks (centralized networks) connect two different types of computers: servers, usually (but not always) powerful computers, dedicated exclusively for running the server applications, and client devices. Clients can be of two main types: thin clients, computers with limited computational abilities or terminals, and fat clients, normal computers which connect to server via special software. Thin clients are specialized computers, their only function is to connect to servers and receive and display some specific sort of information.

    Multi-tier (or n-tier) architecture is used in complex networks for optimizing the network speed and stability. Multi-tier networks can contain, for example, database server, which stores calculation results for faster future access.

    The second classification category, named network topology, describes the way different devices connect to each other.

    Bus or linear topology contains a single line called bus to which all other computers are connected. Each device can broadcast a message to all other computers, when the bus is free. Computers analyze incoming packets and determine if the content of the message was intended for them. Bus networks are simple to build and change and so are commonly used all over the world.

    Mesh topology is when there are at least two devices with two or more paths between them. They are more difficult to design and implement, but their decentralized nature makes them very useful. If every computer in mesh network directly connects to all other computers, it is called the fully connected or complete topology. Such networks are fully decentralized and very stable, but are used very rarely because of the large number of communication links.

    Ring topology is when every computer has exactly two lines connected to it. These computers and lines form a ring. If one part of the ring fails then the ring is broken and cannot work. A dual ring topology has four branches connected to it, and is more resistant to failures. Data in ring networks proceeds in only one direction at a constant speed round the loop. Devices may send information only when they are in control of the “token” — a special package of data. So only one device can send information at the same time.

    In star topology peripheral computers are connected to a central computer, which rebroadcasts all transmissions received from any computer to all other computers on the network, including the originating computer. The failure of a transmission line will result in the isolation of only one computer from all others. Star networks are based on client-server architecture and are quite slow, but are less susceptible to transmission line failures.

    In tree topology all computers are arranged as a tree. This type is based on star topology, but the function of the central computer is distributed between client computers. As a result, tree networks are more susceptible to lines failures, but do not depend of a central computer.

    Switched networks use a central switch to connect different devices on the network directly. Once the link is established, the two devices communicate as though they were directly linked without interference from any other device.

    Most existing computer networks have hybrid topology, which means they are based on several basic topology types. But usually such constructions are based on bus topology.

    The most popular network classification is based on size or range of the network. Using this category networks can be personal area, local area, metropolitan area or wide area.

    Personal area networks (PANs) usually are used to connect several devices close to one person. They can connect, for example, user PDA or notebook to a desktop computer, or a printer, scanner, headphones or any other external device to any type of computer. Such networks can be based on a bus, which can be an USB (universal serial bus) or a Firewire cable. Wireless PANs are based on radio communication and use such protocols as IrDA (an infrared standard, developed by Infrared Data Association) or Bluetooth.

    Local area networks (LANs) cover a local area, like a home, office, small group of buildings, organization, with maximum size not more than 1 km². First standard for LANs called ARCnet was released in 1977 by Datapoint and was intended for sharing disk storage. Today there are two common wiring technologies for a LAN — Ethernet and Token Ring. Ethernet protocol is based on bus network topology and is widely used in public networks. When using this protocol, computers are usually wired to a hub or a switch. LAN can be connected to another LAN or to a WAN with help of routers. The transmission protocol used in LANs is TCP/IP (short for Transmission Control Protocol and Internet Protocol). Each computer has a network address, called the IP address. Several other protocols are used for network control, for example DHCP (Dynamic Host Configuration Protocol) can be used for dynamically acquiring a unique IP address in the network. LANs are mostly used in different organizations to connect office computers, to exchange information and to share organization hardware and software resources. Another case of LAN is home networks, which connect home desktop computers in several buildings. Wireless LANs (WLANs) can be also used to form a LAN between computers on a longer distance. Modern operating systems usually provide easy and convenient tools for setting up and troubleshooting the local area networks.

    Metropolitan Area Networks or MANs are large computer networks usually covering a campus or a city. They typically use optical fiber or wireless connections to link their sites. MANs are usually connected to Internet and sometimes appear when several LANs are joined together. The protocol used in MANs is called Gigabit Ethernet (GbE) and provides connection speeds up to 1 gigabit per second. Such networks based on both optical fiber and wireless technologies cover such big cities as Taiwan, Berlin or New-York.

    A wide area network (WAN) is a computer network covering a wide geographical area, involving large number of computers. WANs are used to connect local area networks together, so that users and computers in one location can communicate with users and computers in other locations. Many WANs are built for one particular organization and are private. For example, IBM Corporation owns a private secure network, which is used by company workers for transmitting important data. The first protocol used in WANs was X.25 superseded by Frame Relay protocol, ATM, IP and many others.

    The world largest public hi-speed wide area network is called Internet. It is based on ARPANET network, created by the United States Department of Defense Advanced Research Projects Agency (ARPA) in 1969. On January 1, 1983, the ARPANET changed its core networking protocols from NCP to TCP/IP, marking the start of the Internet as we know it today. In 1991 the WWW was created. Then the HTTP protocol and the first browser, called Mosaic, were created. The Internet successfully accommodated the majority of previously existing computer networks, connecting computers from all over the world in a largest computer network ever known.

    The Internet is based on several layers of protocols. The TCP/IP, IPX UDP protocols are used for physical packet transmission. The PPP (Point-to-Point Protocol) is commonly used to establish a direct connection between two computers. The DNS (Domain Name Service) protocol is used for acquiring IP address of a computer by its name. Higher level protocols are directly used in services provided by Internet. These include FTP (File Transfer Protocol), POP3 (Post Office Protocol), IMAP (Internet Message Access Protocol), SMTP (Simple Mail Transfer Protocol), HTTP (Hypertext Transfer Protocol), HTTPS (HTTP Secure), Telnet, SSL (Secure Socket Layer), SSH (Secure shell), LDAP (Lightweight Directory Access Protocol) and many others.

    Internet provides to a connected user many convenient services to exchange specific types of information with other users. The most used Internet service is undoubtedly the World Wide Web (WWW), a set of information resources called web pages or web sites, composed using hypertext and identified by URIs (Unique Resource Identifiers). A special computer program name web browser can retrieve pages and present them to a user. The Web was created by Tim Berners-Lee in 1989 on several servers in the European Organization for Nuclear Research (CERN) and since that time has become the world’s largest source of information ever existed. According to a 2002 survey, the Web contained about 2 billions web pages.

    Other popular Internet service, e-mail (or electronic mail) a method of composing, sending, and receiving messages over electronic communication systems. Although e-mail has a longer history than Internet (it started in 1965), most e-mail services now use Internet. The mail can be sent or received by a user with help of e-mail clients using mail protocols (POP3, IMAP and SMTP) or using WWW pages which provide a remote access to your e-mail box with help of browser.

    Internet is also commonly used for file sharing. The FTP (file transfer protocol) is usually used for this purpose. In the beginning of the 20th century the P2P (peer-to-peer) networks appeared, which used special programs to exchange files between the users. The simplicity and convenience of file exchanging over the Internet also creates copyright problems in case of sharing books, music, software and video (clips or movies) protected by copyright problems — or in short piracy. Though several successful processes lead to closing of some popular Internet file exchange services (Napster and ShareReactor, for example), Internet currently remains an uncontrolled community with very indistinct copyright restrictions.

    Some other Internet services are directly used for human communication. These include chat systems, such as IRC (Internet Relay Chat) and instant messaging (IM) systems, such as ICQ (short for “I seek you”) and AIM (AOL Instant Messaging) systems.

    Internet today is the largest source of information and a uniquely fast and easy communication method in the modern world. With the constantly growing role of computers in our life, computer networks slowly replace traditional communication technologies and become a fundamental part of our future.