Networks or Local Area Networks (LAN) were primarily invented to give personal computer (PC) users the same communication and resource sharing capabilities as those of the greater community, ie., the mainframe users, in the early 1970's. These capabilities included and still include sharing of data and resources. LANs were the answer in that these networks allowed sharing of data and resources, while, at the same time, allowing users to retain some autonomy of purely personal computing. 

 Two definitions are in order in this discussion concerning data and resource sharing. Data sharing is straight forward. Digital files created at one computer or workstation on the LAN can be shared with others. Data files can be anything which is digitally generated., documents, spreadsheets, databases, forms, or programs. Each workstation must have similar software application programs with which to open and manipulate the file. The data is moved from workstation to workstation as digital signals. 

Resource sharing, on the other hand, means something a little different. Resources in a network can be either physical devices or virtual devices. Examples of physical devices include printers, hard drives, CD-ROM, modems, or scanners. Virtual storage devices would be an example of software shared devices. LANs provide for the fact that multiple users can share printers for example. Device sharing usually results in overall cost savings to the network owner. 

 The layout aspect of LANs concerns how the network workstations or nodes are connected together. This is called the network topology. In LAN terminology, topology refers to the specific system of interconnection used to link the workstations or nodes together. Workstations and nodes are used interchangeably in describing network topology and are in fact the same thing. The topology refers both to the physical and electronic connection. The physical topology refers to the way the cabling and computers interface the network component connections. The electronic topology refers to the way the data is passed from node to node. 

 Two primary topologies are in use today and are named token ring and Ethernet. A third topology, ARCNET has been used and is not currently being installed very often. Token ring networks are arranged in a ring linked in a single closed loop. Ethernet networks utilize a bus with a common cable to which all nodes are connected. 

 Token ring topologies have been used since the early 1970s. This standard was developed by IBM. The ring hardware calls for special purpose node connectors at each workstation on the closed loop. The protocol requires that data travel in one direction only in a format called a frame. The frame contains the message as well as data which identifies the network address. This topology is extremely complicated. The bandwidth of most token ring networks is 6 to 15 Mbps. 

 Ethernet networks are bi-directional in which signals are transmitted in either direction. The Ethernet bus structure is easy to install and maintain. It is best instituted in a star configuration using HUBs as the relay point for data transmission. Most networks used today operate within the Ethernet standard as described by the IEEE 802.3 standard. Ethernet data rates operate at 10 Mbps or 100 Mbps under today's standards. Ethernet hardware can be purchased off the shelf for the most part. The star topology lends itself to easy maintenance and can be tested for problems node segment by node segment. 

LAN Communication Hardware Components

 Hardware components include the computer and other devices which are connected. Computers do not have to be sophisticated or powerful to be connected. The usual computer hardware components can be utilized such as hard drives, CD-ROMs, scanners, pointers, and modems. One additional component must be included in each computer and is the point where the physical media connects the computer on the LAN. This component is the NIC. The NIC is the physical input/output interface between the computer and the network. The NIC converts the desired digital computer data into serial data format and then transmits it over the network. The receiving NIC then reconverts the serial data into its digital format at the receiving workstation. 

 The computers in the network performs many roles. Some computers function as servers while others function as workstations. Other computers function as specialized systems handling print, file storage, modem or access to other functions. 

Another hardware component required in each LAN is the cabling or LAN media. The most common cabling used is either coaxial cable or twisted-pair cable. Coaxial or coax cable consists of an inner conductor, insulated core, shield and jacket. Coax cable is connected with BNC connectors. Coax cable looks like TV cable and the connectors work like TV cable connectors. 

 Twisted-pair cable consists of one or more twisted solid copper wire pairs. The total number of twisted-pairs may be shielded or unshielded. Most Ethernet cable consists of four twisted-pairs of copper wire in the shielded or unshielded jacket. Most twisted-pair cable is unshielded and is called unshielded twisted pair or UTP cable. Connector terminate each section of the cable to insure a continuous path for the electronic signal. The most common Ethernet UTP connectors are designated as RJ 45 connectors and look like super phone line connectors. The RJ 45 connector has eight terminals to connect each of the eight copper wires. The UTP cable is normally high quality to keep static and other electronic interference signals at a minimum. The cable installed in most networks is rated at Category 5. 

 The final hardware components for many LANs is the HUB or wiring connector. HUBs are nearly always used in Ethernet bus topology networks. HUBs make the network physical arrangement easy to lay out. Most Ethernet LANs are arranged in a star topology in which each workstation node cabling connects into a HUB which becomes the central data distribution point. HUBs can be simple electronic receptacles (passive HUB) or can server as repeaters to amplify or reshape the signals (active HUB). 

LAN Communication Software Components

 Networks do not have to be powerful to have good connections. Windows for Workgroups 3.11, Windows 95/98, UNIX Linux, and IBM OS/2 are network compliant out of the box. Computers running these OS systems need only to install NICs and cabling to be network ready. This connection arrangement is called a peer-to-peer network. File sharing and component sharing are possible in peer-to-peer networks. There is no central capability in per-to-peer networks. Each workstation is equal. The only other requirement in peer-to-peer networks is having compatible software applications such as Word or WordPerfect word processing for document manipulation. Peer-to-peer networks are excellent methods to share printers, CD-ROMs, or scanners, thus cutting the cost of having individual equipment at each computer. This is an excellent way to let home users share equipment. 

 Specialized NOS programs such as Novell NetWare or Windows NT are usually run on server computers. These NOS systems are client/server arrangements and are usually powerful enough to handle all the network requirements. The server waits for requests from the client workstations to perform functions such as network file storage and sharing, printing or other functions. Server based networks provide increased security, controlled access to data, as well as data storage and sharing. 

Conclusion

 Computers which function as servers normally are vary powerful machines with high speed CPUs, large hard drives, and large amounts of random access memory. These are required for the server to be able to handle multiple workstation requests at one time. Servers make secure storage locations for data. By placing all data on the server storage, backup is easily made to the data. 

 The other network hardware components such as cabling, fittings, and connectors work without problems once they are initially setup, tested, and installed. Cable and fittings are readily available in computer stores. Ethernet UTP cable can bd purchased ready to plug into the network. There are some IEEE 802.3 standards limitations which apply such as cable length. A SOHO network would never run into these limitations. 

 HUBs are available "off the shelf" and can be node sized for nearly any network. Ethernet HUBs may have as few as four ports, or up to twenty-four ports. HUBs may be cascaded to add multiples of nodes in expansion programs. 

 The hardware aspects of the SOHO network are easily setup and installed. The benefits to home or small office users are great. 

 JOHN WOODY IS A TELECOMMUNICATIONS CONSULTANT SPECIALIZING IN SMALL BUSINESS COMMUNICATIONS, NETWORKS, AND INTERNET BUSINESS TRAINING.