How fast is fast enough?  The most important element of Local Area Network (LAN) design is how fast you can get data from one point to another on a network. A network designed for 10 MHz (10,000,000 data bits/second) will only transfer one-tenth as much data per second as a network designed for 100 MHz (100,000,000 data bits/second).  A  1 GHz (100,000,000,000 data bits/second) network really screams!

Networks using unshielded twisted pair cable (UTP) originally performed at 10 MHz (Category 3). As more people shared larger networks and transferred larger files, network speed was increased to 100 MHz (Category 5). CAT 5 networks have been with us for some time. In an effort to enhance performance of networks, the components of 100 MHz networks were improved and given the designation CAT 5e. Although both networks run at 100 MHz, CAT 5e is definitely better and more trouble free than a CAT 5 network. Manufacturers have stopped producing CAT 5 in favor of the CAT 5e specification. The minimum network installed today should be CAT 5e.

In an effort to improve speed and reliability, the specification for CAT 5e was again improved to yield CAT 5e+. In general CAT 5e+ will operate up to about 350 MHz.  Category 6 (CAT 6) is the newest standard to be implemented for unshielded twisted pair (UTP) structured wiring systems. The systems supports speeds up to 1000 MHz.

If you need network connections longer than 100M (300') you should be using fiber optic cable. Using multimode fiber optic cables can extend network distances to 1.6 KM.  Fiber optic cable is immune to electromagnetic interference and is the obvious choice if the cable is around heavy machinery or large transformers.  CAT6 and fiber optic connections will both yield speeds up to 1 GHz.

A computer network allows separate computers to share resources within the boundaries of the network. Those resources may be shared files or databases, programs, e-mail access, printers, telephony application processors, FAX capability, or high speed internet access.  Resources may be shared individually by each computer (peer-to-peer networks) or may be controlled by a central computer or server (client-server network.)

Each PC is connected to the network by a Network Interface Card (NIC). Elements of the network may be connected by cable, the internet, dedicated data circuits,  or by wireless network devices.  Selection of network type depends upon speed, distance, cost, and security issues.

A peer-to-peer network allows two or more Personal Computers (PCs) to pool resources.  Individual resources like disk drives, CD-ROM drives, and printers can be made accessible from every PC.  Each PC may act like a “client” and request resources from another PC or act like a “server” and provide resources to another PC. Devices like printer servers, telephony application processors, and high speed internet access can also be accessed via the network.

Network Operating Software (NOS), usually contained in the PC operating system, allows each PC to determine which resources will be available to other users. Specific hard & floppy disk drives, directories and files, printers, and other resources can be attached or detached from the network via software.

In a Client-Server network, files are stored on a centralized high speed file server. Network access speeds are usually higher than speed on a peer-to-peer network. Nearly all services such printing, internet access,  and e-mail are routed through the server, allowing them to be tracked. Client PCs are subordinate to the Server and have their capabilities, settings, and access to resources set by programming on the server.  A Client-Server network centralizes administration and simplifies administration on large networks.

Network printing uses a small print server (or a print server card built into the printer) connected directly to the network. Printers can be used on a peer-to-peer network without having to turn on the PC with the printer attached. Printing is faster and less cumbersome. Printer servers may be used on either peer-to-peer or client-server networks.

10BaseT and 100BaseTX are the two most common standards for networking speed. 10BaseT was the first standard for unshielded twisted pair (UTP) cable for networks. It operated at a speed of 10 Mbps (10 million bits per second) and used two twisted pairs of wire. 100Base TX operates at a speed of 100 Mbps (10 time faster than 10BaseT) and uses four twisted pairs of wire. New standards of 350 Mbps and 1000 Mbps have been approved and are now being implemented.  Installation costs for the new standards are higher.

Auto sensing refers to network devices like hubs, switches, and network adapters which automatically sense which speed the network is using and adjust to the proper speed. Current equipment is primarily designed for 10/100 speed. New auto sensing equipment will eventually be marketed for the higher speed networks.

CAT 3 and CAT 5 are common standards which define the cable used for networks.  CAT 3 was designed for 10BaseT networks which operate at 10 Mbps. CAT 5 was designed for 100BaseTX networks.  In addition to electronic and physical specifications for the cable, the standards define specifications for patch panels, jack inserts, patch cords, network adapter jacks, and installation practices.  Also defined are a CAT 5e + (155 Mbps) and CAT 6 (250 Mbps). CAT 6 has recently been approved and the equipment and testers are not yet fully available.

Full Duplex devices send and receive data simultaneously - a full duplex switch sending and receiving data at 100 Mbps is actually transferring data at 200 Mbps. A Half Duplex device can only send or receive - not both - at one time. Switches are usually full duplex devices and have greater throughput of data. Hubs are half duplex devices and are more limited in throughput.  Full Duplex devices are more expensive than half duplex devices.

In wireless networking, cables are replaced by wireless transmitters and receivers and may be either full duplex or half duplex. Wireless devices may be difficult to integrate into a wired network and there may be some security and coverage issues that need to be resolved.  Wireless networking is probably more useful for a very small network.

Hubs and Switches are devices which connect computers in a network and help transfer data across the network. Hubs are either 10 Mbps or 100 Mbps and are usually half duplex.  Computers on a hub share the total bandwidth available to the hub. If you have a 10 Mbps 5 port hub, each of the five connected PCs has available only 2 Mbps of bandwidth.

Switches use dedicated bandwidth for each port. A 10 Mbps 5 port switch will provide each device with a full 10 Mbps data speed.  In addition, since switches are usually full duplex, the actual data throughput may be as high as 20 Mbps. Switches are more expensive than hubs, but increased performance probably justifies the difference in cost.

Routers provide two valuable network functions:  They act as traffic cops for network traffic making sure data gets to the right place and they provide and interface or gateway to other networks and the internet.  Routers have programmable software which keeps track of devices on the network and guides data packets to the correct device.  They also keep track of incoming and outgoing data from external networks (a company wide area network or the internet) and routes the data to the correct network PC.  They may have additional software to provide protection to network computers (a firewall) and enhanced IP addressing via DHCP (see below).

Using Internet Protocol (IP), each computer and each networking device is assigned an identifying number referred to as an IP address. This address is in the form of XXX.XXX.XXX.XXX and must be unique for every device on the network.

Static IP addresses may be assigned to each network device. These numbers are permanent and the network administrator must keep track of all IP addresses to make sure there are no duplicates and addresses are assigned to proper devices.

For networks with large numbers of computers and with people constantly logging in and out a better method of assigning IP addresses was needed. Dynamic Host Configuration Protocol (DHCP) actually “leases,” or temporarily assigns, IP addresses as computers log on to the network.  The address belongs to the computer for the duration of the session but goes back to the pool of available addresses as soon as the computer logs out.  The process is completely automatic and does not require the record keeping characteristic of static IP addresses.  Use of DHCP by routers simplifies set up of an internet sharing network.