Coupling those two or more computers together requires several pieces of hardware and software that are not commonly part of the basic PC. First, the hardware components include at a minimum, cabling, NIC (Network Interface Cards), and possibly hubs. Second, the OS (Operating System) must have utilities that recognize other computers, to make the OS “network ready”. This overview will attempt to cover the how and what of networks.

I have been writing about SOHO (Small Office, Home Office) networks since 1997 in my Comm Corner column. In these Comm Corner articles about SOHO, I defined it as a network when two or more computers were connected together. There is no real difference between two or more computers connected together at home or in an office. 

Communication by most of us in the beginning, 1980's or so, was via word of mouth, passing a floppy disk to one another or using our analog modem to connect to the local BBS (Bulletin Board Service) or another computer user at 300 baud on the POTS (Plain Old Telephone Service). The Microsoft DOS OS was not network ready without real black magic additions to make it recognize the network hardware and software. Other OS systems, were, at this time, capable of networking or providing the means of connecting the computers together. UNIX, being used by universities, business, and government had connecting capabilities. Most networks were the prerogative of the major corporations, who used proprietary OS on really expensive computers. Novell and Banyan were two of these. The hardware portion of these networks were big and expensive, as was the OS. The installation and maintenance was technical and expensive as well. There was, however, a networking OS protocol running in the background, that would become the basis for today’s network capability. More on this later.

The home or small business user was left in the cold. Microsoft, in it’s follow-on OS systems began to provide various levels of network capability. The PC world that most of us knew had to wait for Windows 3.x including Windows for Workgroups and the Windows 95/98/ME OS series to gain this network capability. Microsoft developed a series of business OS applications that are really network ready in the later part of this period. NT/2000/XP OS systems fall into this category. The Apple OS systems had network capability in this time frame. The hardware side of the network equation, NIC, hubs, switches, and routers began to be built and priced so that the home and small business user could afford to install networks and leverage the capability of several PCs. Computer technology also developed so that the power of the individual computer could match early mainframes.

Also, as the technology of the Internet and other large scale networks became available, that technology fed into the PC, making it a real communication tool. Advances in connections to the Internet, i.e., broadband telco DSL and cable television Direct Cable moved all of us away from analog modem connections and the speed limitations of the analog telephone system. The increased bandwidth of broadband made it possible for more than one computer at a single location to be connected to the outside. This multiple use communication capability literally propelled the home and small business user into networking. The broadband bandwidth gave individual users the same multiple use capability that big business had in the early computing where they used leased lines and expensive T-1 connections. Individual broadband also drove the networking components (routers, NICs, hubs, and switches, etc.) prices down. Manufacturers have been able to bring the price of network components to within everyone’s reach due to the large numbers of individual and small office users.

Basic Network Technology
The basic physical and concept layout of networks include the medium (means of data transfer, i.e., cable, radio transmission, etc.), access control method (OS protocol), transmission technique (broadband, or baseband) data rate, and network operating system. The primary function of networks is to provide data and resource sharing capabilities to everyone on the network.

Data and resource sharing includes two definitions that must be understood to really understand networks. Data sharing is straight forward. Data can be shared in any number of ways including copying to a floppy and hand-carrying it to the intended user, or it can be transmitted electronic means, i.e., the network. Digital files created at one computer can be shared with other computers on the network. Digital files can be anything that is digitally generated, documents, spreadsheets, databases, forms, applications, or graphics. 

Resource sharing is a little different in that a resource in computer terms is anything that is available in the computer or network. A resource can be a physical device such as a printer. It can also be a virtual device such as a virtual storage capability. We mainly think of physical devices such as hard drives, printers, CD-ROM, modems or scanners. The shared resource on the network such as a printer can usually results in an overall cost savings to the network owner in that everyone on the network can use that shared printer or other resource. Data safety is also enhanced using a network to store the data on more than one location.

The means of data transfer or media refers to the method of data communication. This media may be a cable, radio frequency, or other electronic method of sending and receiving the data. The cable may be a direct connection from one serial port to another via a null cable. The cable may be high quality twisted pairs that meet certain standards such as Category 5 Ethernet cable or it may be coaxial like TV cable. Fiber optic cable is used in very high speed data transfer situations. And, the data can be transmitted and received by radio frequency network methods.

The access control method is found in the OS protocol that controls how data moves through the computer. Within the OS, a structured standard is included called the OSI Model (Open Systems Interconnection) that consists of seven layers that perform how the data will flow through the computer and network. All network ready OS systems must conform to part or all of this model. Each layer of the model is interconnected to the layer on each side if it to handle how the data is addressed and packaged for transmission and receipt. Each layer specifies how the hardware and software interacts also. Every computer or node in the network must recognize and utilize the same OS protocol.

The transmission technique governs how the data is handled as it flows from sender to receiver. The network standard specifies how fast and in what form. An Ethernet network, for example, specifies what the hardware and software shall do at each stage of the data transmission. Ethernet NICs, cable, and other devices are required. The network OS protocols handle and prepare the data to meet the Ethernet processes.

The standards specify how the network is arranged. The layout aspect of the network concerns how the network computers or nodes are connected. This is called a network topology. The topology refers to the physical and electronic connections. An Ethernet network topology, for example, is usually configured in a star arrangement. All of the nodes or hosts (computers) in the Ethernet network are connected to a central point, usually a hub or switch, which acts as the traffic director for the data flow. Other topologies include token ring and fiber networks that are arranged in rings in which the data flow travels in a ring to its destination. The Ethernet, for example, standard is described by IEEE 802.3 and covers how the data is handled at each stage of its movement as well as what the data rates will be. Ethernet data rates are from 10 Mbps to 100 Mbps. The IEEE 802.3 standard is being modified to let the data rates move to Gbps.

The home and small office network can expect to operate at 100 Mbps within it’s LAN. It’s outside connection is governed by the speed of the outside network, i.e., DSL at 1145 Kbps downstream/ 142 Kbps upstream or Cable at 2723 Kbps downstream/ 332 Kbps upstream. (Data from actual tests on my two connections during a previous test. April 2002 Comm Corner column.) The combination of bandwidth and speed that is available to the home and small business user is the basic factor that makes a home or small business network feasible. The bandwidth can be shared by all the computers in the network.

Network Hardware Components
Hardware requirements in addition to the computer include the NIC, and possibly a hub or switch. In the beginning, the NIC was a add-on card to either a ISA or PCI slot on the MB (MotherBoard). Now the NIC can be a built-in port on the MB. The Ethernet IEEE 802.3 standard specifies how the NIC is designed and what protocol it will recognize from the OS. The plug-in connector must also be specified for the unit to work. The most common Ethernet female connector ports are designated as RJ 45 and can receive eight wires from the cable plug. They look like a very large telephone connector. The NIC is specified as 10BaseT or 10/100BaseT. This indicates the bps (bits per second) the unit will transmit or receive data.

Ethernet cabling is specific and requires four twisted-pair cable, either shielded or unshielded. Shielded cable is very expensive, so the most common cable is unshielded and is called UTP (Unshielded Twisted Pair). The cable is of high manufactured quality known as Category 5. The cable is terminated at each end with RJ 45 plugs so that the eight wires are arranged in a specific order. UTP Cat 5 cable will transmit or receive at up to 100 Mbps.

A companion part of the IEEE 802 Ethernet series standard allows for radio frequency as the transmission medium. This part of the standard is known as IEEE 802.11 series a or b. The 802.11b standard transmits and receives at up to 10 Mbps.

The next two pieces of hardware that can be required are Ethernet 10/100 hubs or switches. A hub is a device that directs the computer traffic within the network. A switch is a smart hub. The cable from each of the computers is plugged into a receiver connection in the hub, which then redirects the traffic from the sender to all the receivers. The switch knows which computer to send the traffic to and directs traffic only to the proper receiver.

One additional hardware component in the network, especially with broadband an outside connection is a router. A router is a device that knows what addresses are within the network and what address is at the other end of the broadband connection and can direct traffic to the proper receiver. Most Ethernet broadband routers have a built-in switch, either two or four RJ 45 ports. 

Broadband routers also provide a level of security to the network. Broadband connections, DSL or Direct Cable, are open all the time the computer or router is running and is connected. This can be a source of unauthorized entry into the computer or network. The broadband router has built-in utilities that protect against this unauthorized entry in the form of IP (Internet Protocols). Two IP utilities are available in most of these routers. IP NAT (Network Address Translation) and DHCP (Dynamic Host Configuration Protocol). NAT provides a single IP address to the outside world while allowing multiple IP addresses inside the network. DHCP is a service that assigns the internal IP addresses to all the computers or hosts inside the network. The network ready OS has the ability to assimilate the IP address assigned to it from the DHCP server service in the router.

Network Software Components
In order for an OS to be network ready, it must have the OSI utilities that are referred to as protocols. These protocols direct the data handling at each stage of the transmission and receipt. Addressing and packet handling, the technique of breaking the data into usable byte sizes for transmission and receipt, are done by the protocol. The default standard that is universally used today is the TCP/IP (Transmission Control Protocol/Internet Protocol) standard developed for the Internet.  This protocol is included in nearly all OSs being used today. Windows 95/98/ME/NT/2000/XP are equipped with the TCP/IP protocol suite. Apple, Novell, UNIX, and Linux also contain this protocol suite. This universality use of this protocol means that all computers containing it can talk to one another regardless of the basic OS. The IEEE Ethernet standards recognize this IP protocol suite.

The IP protocol suite is based on a number scheme for addressing each host in the network. This addressing scheme makes each host unique as it is the only address in the network. Each host and the router know who is on-line at any given time. Some addresses are Internet ready and can be recognized by any computer that is on the Internet. Others, used primarily in the DHCP portion of the router service are not readable on the Internet, but are readable within the local network.. This is part of the security set in place by the router. The NAT service is able to translate the outside Internet ready IP address into the proper internal IP address. 

The OS that is network ready can be used in simple or complex networks. An example of a simple network would be two or more computers so that each computer is equal to the next. This is a peer-to-peer network and provides for file and print sharing. Each computer or host in the network must know the other computers. If there is an addition, then each host must update its addressing scheme to recognize the new host.

Other networks operate as server/client nodes in the network. The network OS in this situation operates either as a server, in that the computer running the server OS waits for other computers, clients, to ask it to do things. These functions are varied, file storage, printing, security, DHCP, e-mail exchange, virus protection, firewalls, application service, and database warehousing to name a few services. The client hosts in the network, on the other hand, function as the user hosts to do word processing, spreadsheet handling, graphics, accounting, gaming, and just about any other function that is required by the users. OS systems that fall readily into this category are the Windows NT/2000/XP series, UNIX, Linux, and the proprietary network OSs such as Novell.

Advantages and Disadvantages of Networks
The major advantage of using a network is that the network leverages user’s ability to communicate with other users. The network provides efficiencies in that each computer does not have to have a printer to print, large storage for data, backup capability, or a modem to communicate. A shared printer is available to all the users on the network. It can be connected to one computer or can be node on the network as a network ready printer. That one printer or like device can handle the work of all the users on the network. The storage of data can be installed on more than one computer or server so that there is less chance of losing it. Failures do happen and there is better chance of recovery if the data is located throughout the network. Using a server as the data storage location, only one machine must be backed up as all the data can be stored on it. The increased bandwidth afforded broadband connections coupled with a network allows more than one user to surf the Net or send data or e-mail.

Disadvantages of networks include greater complexity, more chance for problems to occur, and some increased cost in the initial installation as the network components must be accounted for. Networks can be complex and need some technical expertise to setup and operate. Problems do occur. The network can crash. More components in each machine mean that more can go wrong.

In the final analysis, the advantages far outweight the disadvantages. The leverage gained by networking greatly increases the communication that users may exercise.