As with most complex technologies, there's no one-size-fits-all solution when it comes to networking. The needs and resources of each unique organization will correlate to a different set of networking solutions. This means that an individual or team that knows their organization and the relevant options will need to carefully consider the situation and determine the optimal network design for their situation. In this, the second article in our Networks 101 series, we'll discuss the major decision points to consider when planning a new network and the questions you should use as the basis for your decisions. See Networks 101: Concepts and Definitions and Networks 101: An Introduction to Server Applications for more.

The following are some questions that might help to guide your network planning process:

• What networking infrastructure do you already have in place?
• How many computers and networked devices do you have?
• Does your office's design and layout impose physical constraints on your planning process? For example, is there available space in your floors, walls, or ceilings (often in the form of electrical conduits) where you can string your network cables?
• What networked applications do end users rely on most heavily, and how much bandwidth do these applications consume?
• Are you planning any changes to your technology infrastructure (such as additional employees or new applications) that might have an impact on your networking needs?
• How much money do you have budgeted for the installation and maintenance of your networks?

Networking Options

The answers to the questions above will determine how your organization designs and builds its network. The first determination is deciding what scale network you're trying to design or re-design. The vendors you'll work with, the technologies at your disposal, and the decisions you'll have to make vary considerably depending on whether you're building WAN links or LAN infrastructure. Read on to learn what those terms mean.

Network Size and Scale

Scale is usually the first and most important determination in network planning as it will often determine or influence your other decisions. While there are other scales discussed in networking literature, most nonprofits and libraries only need to focus on the two or three network scales that impact small and mid-sized organizations (LANs, WANs, and CANs).

A local area network (or LAN) is designed and implemented at the scale of a single building or office. Its primary function is the interconnection of the computing resources within a single organization. In most cases, LANs use Ethernet over twisted-pair cabling or wireless technology.

A wide area network (or WAN) connects a single office or branch LAN to its parent organization's network and all the millions of networks that together make up the Internet. Most authorities define WAN as a network that crosses one or more jurisdictional boundaries (metropolitan, regional, or national). WAN links usually fall under the purview of ISPs and telecommunications companies. Very few organizations have the resources to build and maintain their own WAN links, and it's usually more cost-effective to lease them from the local phone or cable company. WAN links depend on numerous technologies that vary considerably in terms of speed, cost, and bandwidth. For more information, see Choosing the Best Internet Connection and Wikipedia's article on WANs.

A campus area network (or CAN) connects multiple LANs belonging to the same organization when they're in close geographic proximity. Since most nonprofits operate at the scale of a single office or main office with branches, CANs are only a consideration for the largest organizations. As with WANs, an organization can build its own CAN, but most opt to lease facilities from their local ISPs.

Network Cabling and Other Hardware Options

All computing relies on physical components and operations at a fundamental level, and networking is no exception. As with other aspects of computer networking, there are a bewildering variety of network hardware devices available and a lot of terms and concepts to keep track of. You can ignore most of these until you're ready to plan your network design and equipment purchases, but a handful of definitions and distinctions are part of the assumed background knowledge in most networking conversations. Also, some basic knowledge is required in order to make smart, effective purchasing decisions.

The following are some issues you should consider when planning purchases of networking hardware.

Data cables (also known as transmission media) are responsible for carrying messages back and forth between computers and other devices and as such are the foundation of your network. All other network equipment has to be compatible with your choice of data cable, so this decision constrains or determines many of your other choices. While dozens of cable variants are standardized and available for purchase, these variants fall into four main categories:

1. coaxial cable
2. twisted pair cables
3. optical fiber
4. wireless

Of these four, twisted pair cables or wireless transmission are the basis of most LANs. Coaxial cables and fiber optic cables can operate at higher speeds, but they’re more expensive and harder to install, so they’re used primarily by well-resourced organizations that need the high-speed, high-bandwidth networks.

The standards for transmission media are set by engineering organizations such as the IEEE and change on a regular basis, but currently the most common types of twisted pair cabling (in order from slower to faster) are Cat 5, Cat 5e, and Cat 6. The latest, fastest wireless standard as of late 2010 is 802.11n. For more details, read the Data Cabling Chapter of An Educator’s Guide to School Networks.

Equipment Types and Definitions

Networking equipment has a set of functions similar to those of the United States Post Service. Receiving thousands or millions of messages each day, both have to send the message to the correct destination or at least forward it to another messenger who's closer to the recipient. For these systems to work properly, every entity expecting to send or receive messages must have a unique address. The most commonly-used, essential pieces of networking equipment keep track of nearby addresses and send electronic messages (also known as data packets) closer to their destinations.

The role of the local mailman in computer networks is usually played by a device called a switch (also known as a network switch or an Ethernet switch) which distributes messages among the computers and other devices on a single LAN segment. Cisco is the leading manufacturer of switches and other networking equipment, and they donate several of their switches to eligible 501(c)(3) nonprofits through TechSoup. For a longer definition of a switch and a description of how they work, read How LAN Switches Work on Cisco’s site.

If you're sending a message beyond your neighborhood, there are too many addresses in the world for a single switch to keep track of. Routers are comparable to the post office's regional distribution centers. Any message with a non-local address goes to the router which is responsible for forwarding it to another router closer to the final destination. For a longer description of router functionality and mechanics, see Cisco's Routing Basics article. In addition to the switches mentioned above, Cisco also donates a selection of routers to eligible organizations.

Furthermore, just as every entity partaking of the post office's service needs a mailbox or P.O. box, every new computer has a network card (also known as a NIC) that sends and receives messages on behalf of the operating system and applications using that machine. The size, shape, and engineering specifications of network cards vary depending on the type of network involved. For example, an Ethernet NIC will have a port on the back that accepts a standard RJ45 connector, and a wireless NIC will have an antenna on it for sending and receiving radio waves. Regardless of type, manufacturers burn a unique number known as a MAC address into each network card.

Knowing Your Address

Finally, just as the postal system uses two numerical addressing schemes (street addresses and zip codes) for identifying recipients and routing efficiently, networks and networking equipment rely on two sets of addresses to route messages quickly and efficiently. A detailed understanding of these addressing protocols is something usually delegated to IT staff or consultants, but a few of them are so important to the maintenance of a functioning network that it's good to know where to look for help in case your experts are unavailable. If you need to know more about MAC addressing, read An Introduction to MAC Addressing and How to Find the MAC Address of Your Computer. IP addresses are the other major form of machine identification on the modern Internet. If you need to find your machine's IP address and understand its structure and meaning, consult Cisco's IP Addressing and Subnetting for New Users or Understanding IP Addressing and Subnetting on Microsoft Technet.

Conclusion

Though many smaller nonprofits or libraries may hire an external contractor or IT support person to help them build and develop effective networks, understanding the basic roadmap and definitions for building a network can help ensure that the results work best for their needs. Spend time trying to accurately answer the questions listed at the start of this article and weigh the options listed in the sections when deciding on the network that will suit your organization’s budget, IT staff, and needs most appropriately.