Networking 101a

A Beginner's Guide To Networking

Networking, at its most basic is simply the act of linking two or more computers together. While there are many, the most common types of networks, and the ones we'll be focusing on are LANs (Local-Area Networks), and WANs (Wide-Area Networks). LANs are confined to a local area such as an office building, while WANs can be accessed over distance via telephone, radio or satellite link. Networks are generally defined by their architecture (peer to peer or client/server), topography (the geometric wiring structure in which the computers are linked), and protocol (the specific method in which data is passed between units). We'll delve deeper into these properties as we progress. In Networking 101a we will concentrate on typical "hard-wired" LAN techniques and methods. Future articles in this series will cover HomePNA (a method of building a small network that uses in-place telephone lines as the medium for moving data) and local "wireless" networking techniques.

LANs today are powerful, flexible, and easy to use networks that incorporate many sophisticated technologies all working together in harmony. They allow users to share computer resources electronically, including hardware such as printers and CD-ROM drives, software applications, and, most importantly, the data users need to get their work done. Building a LAN is a process of choosing and assembling a number of types of hardware and software and matching them together. In order to accomplish this it is necessary to understand the fundamentals of how LANs operate. What different technology choices are available? Why should you choose one option over another?

bus.gif (2292 bytes)

Today's LAN is shared access technology, meaning that each of the devices attached to the LAN share a single communications medium, usually a UTP (unshielded twisted pair), coaxial, or fiber optic cable. In the figure above, several computers are connected to a single cable that serves as a communications medium for all of them. Physical connection to this medium is made by means of a network interface card (NIC) that plugs into either a PCI or ISA slot inside the computer and connects to the network cable. Once the physical connection is made, networking software is used to manage communications between each of the computers or stations on the network.

In a shared media network, when one station wishes to send data to another station it uses the software in the workstation to put the data into an "electronic envelope." This envelope, called a packet, consists of the original data along with a header and trailer that carry instructions used by the network software to the designated station. One of the pieces of information placed in the packet header is the address of the designated station.
The NIC transmits this packet onto the LAN. The packet is transmitted as a stream of data bits (fluctuating electric signals). While it travels along the shared cable, each of the stations on the network see the packet. The NIC in each of the stations checks the destination address in the packet header to determine if the packet is addressed to it. When the packet reaches its destination, the NIC at that station copies the packet and then takes the data out of its "electronic envelope" and gives it to the computer.

If the data to be sent is too large to fit into one packet, it will be sent in a series of packets. With shared access, many stations can all share the same cable. Since individual packet size is quite small, it takes very little time or resources to travel to the ends of the cable. So after a packet carrying a data between two stations is transferred, another station can begin transferring data to whatever station it needs to.


star_hub.gif (3148 bytes) The most widely used LAN technology in use today is Ethernet, and it is the network solution we recommend for any but the smallest home network. It strikes a excellent balance between speed, price, ease of installation, and support. Ethernet networks run at a data rate of 10 million bits per second (10 Mbps) or, in the case of Fast Ethernet, 100 million bits per second (100 Mbps). This is according to standards set by the Institute of Electrical and Electronics Engineers (IEEE) in a specification commonly known as IEEE 802.3.

The access protocol used by stations connected to an Ethernet LAN is called carrier sense multiple access with collision detection (CSMA/CD). According this protocol, each station contends for access to the shared medium. If two stations try sending packets at the same time, resulting in a collision on the LAN, the CSMA/CD access protocol is designed to correct and restore the network to normal activity quickly.

Ethernet networks can be configured in either a star or bus topology and installed using any of three different media. Coaxial cable was the original LAN medium using bus topology. Bus topology is rarely used anymore as it is relatively difficult to add new users or move existing users from one location to another. It is also really difficult to troubleshoot problems on a bus LAN with more than just a few stations. Today, most new LAN installations use star topology. In a star topology, each station is connected to a central hub by an individual length of shielded or unshielded twisted pair(STP or UTP), coaxial or fiber optic cable. By far the most common is twisted pair because it is inexpensive, and very easy to install, troubleshoot, and repair. The cable is connected to the station's NIC at one end and to a port on the hub at the other.

Of the two types of twisted pair cable, UTP is the most commonly used. The specification for running Ethernet on UTP is called 10BASE-T. This stands for 10Mbps, baseband signaling (the signaling protocol for Ethernet networks), over twisted pair cable. Other Ethernet specifications include 10BASE-F, a standard which allows Ethernet to run on fiber optic cable, 10BASE5, which uses a heavy coaxial cable, and 10BASE2, which uses a fine coaxial cable. 10BASE5 is seldom used in new Ethernet networks anymore, and 10BASE2 is used only in very small office networks.

Fast Ethernet
Fast Ethernet is simply an extension of the 10BASE-T Ethernet standard, retaining the CSMA/CD protocol, Fast Ethernet transports data at 100 Mbps. Two types of Fast Ethernet are currently in use: 100BASE-TX, which runs over Cat. 5 UTP, and 100BASE-FX, which operates over multimode fiber optic cable.

Token Ring

ring.gif (3839 bytes) Another LAN technology in use today, generally on larger scale LANs is IBM's Token Ring. Token Ring networks can be run at two different data rates, 4 Mbps or 16 Mbps.
The access protocol used on Token Ring networks is called token passing. Token passing is an access protocol by which collisions are prevented by allowing only one station to transmit at any given time. This is accomplished by passing a special packet called a token from one station to another around a ring

A station can only send a packet when it has the token. When a station gets the token it transmits a packet which, unlike the bus topology, travels in one direction only around the ring, passing all of the other stations along the way. This packet is usually addressed to a single station, and when it reaches that station the packet is copied. The packet then continues to travel around the ring until it returns to the sending station, which removes it passes the token to the next station on the ring.

Fiber Distributed Data Interface
Fiber Distributed Data Interface, commonly referred to as FDDI, moves data at a much higher transfer rate than either Ethernet or Token Ring: 100Mbps. FDDI networks used to require fiber optic cable, but today they can also be run on UTP. Fiber optic cable is still preferred on most FDDI networks because it can carry data over greater distances than UTP cable. Like Token Ring, FDDI uses a token passing protocol. FDDI is the preferred method for link several smaller LANs within a large corporate structure.


hubs.jpg (3511 bytes) Ethernet networks require a Hub (Token Ring hubs are generally referred to as MAUs). The hub is a hardware device that each station on the network is connected to by cabling. The hub manages the transfer of data between networked devices. Hubs come in many different port configurations (4, 8, 12, 24, etc.) depending on the number of PCs you want to connect together.

On a simple ethernet LAN each port supports a single 10Base-T connection from a PC workstation or peripheral; such as a tape backup drive or printer. Many hubs available today can also support both 10base-T and Fast Ethernet and are referred to as 10/100 hubs. Aside from it's port configuration hubs come in two styles managed and unmanaged. Managed or Intelligent hubs are hubs that can be monitored and managed through software applications by network administrators. Intelligent hubs make troubleshooting and fault isolation easier and faster because each station attached to the network can be monitored individually and, if necessary, can be easily turned off.

A typical ethernet hub has multiple user ports to which computers and peripheral devices such as servers are attached. Each port supports a single 10BASE-T twisted pair connection from a network station. When a data packet is transmitted to the hub by one station, it is repeated, or copied, onto all of the other ports on the hub. In this way, all of the stations "see" every packet just as they do on a bus network, so even though each station is connected to the hub with its own dedicated twisted pair cable, a hub-based ethernet network is still a shared media LAN.

Managed Hubs
On a managed each of the ports on the hub can be configured, monitored, enabled, or disabled by a network administrator from a management console. Managed hubs can also gather information on a variety of network parameters, such as packet volume through the hub and each of its ports, what types of packets they are, whether the packets contain errors, and how many collisions occur. Hub vendors bundle various hub management applications along with their products. These applications are quite diverse in their methods and parameters and generally you pay more for increased options. For the most part managed hubs are unnecessary for small LANs and we won't be covering them in any kind of depth. Our primary focus will be on standalone unmanaged hubs.

Standalone Hubs
Ethernet hubs and Token Ring MAUs come in three basic configurations: standalone hubs, modular hubs, and stackable hubs. Standalone hubs, as the name implies, are single unit hubs with a fixed number of ports. Most offer some means of linking to other standalone hubs, and are the most cost-effective type of hub. They are best suited for small business, individual workgroups, or home offices.

Modular Hubs
Modular hubs find their niche in larger business networks, due to the fact that they are easily expanded and always have a management option. Beginning with a chassis, called a card cage, with several card slots, modular hubs can be built on the fly, as it were. Each card slot accepts a communications card or module that acts like a standalone hub, which then connects to a communications backbone linking them all together so that a station connected to a port on one module can communicate with a station on another module.

Stackable Hubs

Similar to standalone hubs except that several of them can be stacked or connected together. When linked, they act like a modular hub in that they can be managed as a single unit. By incorporating one managed hub within a stack, you can provide management capabilities for each of the other hubs in the stack. Stackable hubs are best suited for use by medium sized businesses or organizations as they require a minimal investment but can easily grow as the company flourishes. stacks.jpg (3555 bytes)

Enterprise Networking
Bridges, Routers & Switches


MediaTek All Rights Reserved