The first step in developing a network plan is to understand the requirements of the organization that will run the network and the needs of the people who will use it. After you have determined the company's requirements and the users' needs, you can set about developing a technical configuration that satisfies those requirements and needs.
The first order of business is to determine why the organization wants a new network installed. You might be providing a network for a brand new installation that has no existing equipment, or you might be asked to network a group of existing stand-alone computers. A third possibility is that the organization has an existing network that it wants to upgrade to a new or different technology. Each of these situations presents a different set of user and administrator requirements.
An organization with an existing network (or a group of computers that it wants to connect to a network) probably has some idea of what services it wants the network to provide, whereas an organization wanting a new installation might have less of an idea about what it needs. At the most basic level, businesses usually want their users to be able to access shared drives and printers and connect to the Internet. These factors alone allow the administrator to make some basic decisions. Sharing drives and printers is a fundamental networking task that almost any network technology can provide, but you know that the computers will have to run the TCP/IP protocols to access the Internet and that you will need a router of some kind to connect the network to an Internet service provider (ISP).
Some organizations might have more elaborate requirements. It is also important to remember that when it comes to computers, businesspeople often know what they want but have no idea what's involved in getting it. For example, some organizations might need to maintain elaborate databases, which requires powerful servers with fast processors, a lot of memory, and large hard drives. Other companies might need to work with huge graphic images or full-motion video, which requires enormous amounts of network bandwidth and a lot of storage. A financial company might want to provide its users with real-time data feeds for multiple stock markets around the world, which requires large amounts of internal bandwidth and high-speed Internet access. You need to determine what the organization wants and what they can afford, and then try to find a solution that satisfies both requirements.
In addition to talking to the heads of organizations about their network needs, it's also a good idea to spend some time in the trenches with the people who will actually be using the network. You might discover things that can affect what equipment you select for the network, where components should be located, and what ergonomic factors you should consider. For example, the owner of a company might decide that three laser printers are sufficient to service all of the network's users, but the users will be able to point out the best locations for the printers and explain that a printer with an envelope feeder would save a lot of time otherwise spent manually feeding envelopes.
The next step in developing a network plan is to examine the site where you plan to install the network and assess the conditions in which the equipment will operate. This site evaluation helps you to determine which basic network type you should choose, which network medium you should use, and how you should install it. With this information, you can begin to consider the requirements of your users and start to choose appropriate hardware products. Some factors that you should consider when evaluating the site are covered in the following sections.
One of the most crucial elements of the site evaluation is an understanding of where the various computers and other network components must be located in relation to each other. On a typical 10Base-T or 100Base-TX Ethernet network, computers can be up to 100 meters away from the hub, which is far more than necessary in the average office networking environment. However, you should be sure to consider the actual route that your cables will take, not just the distance between the components. Cables must often snake up through a wall, through ceilings, around lighting fixtures, around doorways and other obstacles, and back down through another wall to complete a connection. The final length of the cable can be much longer than just the walking distance from a computer to the hub. You'll need this information to know how much cable you'll need. Or, if you're going to use prefabricated cables, you'll need to know how long they have to be. In other cases, you might find yourself connecting computers that are great distances apart, on different floors, or even in different buildings. In these circumstances, you might need to consider a fiber optic networking solution, which can span longer distances and run safely outdoors.
Another consideration is obstacles between the computers that are to be connected to the network. This factor can determine what kind of cabling job is needed for the site. For example, if the network will consist of computers that are all located in a single room, you might be able to network them using prefabricated cables running loose around the perimeter of the room. This is a relatively simple project that requires no special equipment. However, if the network will be a large one, if the computers are located in many different rooms, or if the site requires the most professional appearance possible, you have to plan an internal cable installation. This type of installation uses bulk cable, installed into ceilings and walls, connected to a central patch panel at one end, and connected to individual wall plates at each of the computer locations. Internal installations require additional planning, more equipment, and greater expertise. They also add significantly to the cost of the job.
All network plans require an evaluation of the areas where the workstations and other client components will go, but on more elaborate network installations you must also consider where to put the back-end components, such as hubs, servers, routers, and so on. In all but the smallest installation, it is important to secure this type of equipment physically, such as in a locked room or closet, so that it's protected from damage, either accidental or malicious. It's not uncommon for expensive tower servers located in public areas to be turned into stands for potted plants or targets for refrigerator magnets, or even to disappear. Depending on the size of the installation, you might need several locking closets in which to put servers, hubs, and patch panels; you might even need a full-fledged data center that contains all of these things and functions as the cabling nexus for an entire building or campus.
The science of ergonomics involves more than the development of strangely shaped keyboards. Part of the job of planning a network includes selecting equipment designs that are suitable for the working environment and placing them in locations where they will provide the most utility and cause the least distraction. For example, you might be able to make one user very happy by moving the departmental laser printer off her desk and onto a printer stand in a more central location. This might mean spending a few extra dollars on an external print server device that enables you to connect the printer directly to the network cable, but if more people have more convenient access to the printer, it's probably worth it.
Another factor to consider is the selection and placement of client workstations themselves. Many of the computers designed for office use have small-footprint cases, which are either desktops or minitowers. Even these smaller cases can occupy a lot of space on the average user's desk, and you might want to avoid this by purchasing computers that can easily be placed on the floor or that fit under or beside desks.
In addition to the physical locations of the network components, you must also be aware of the physical environment in which they will run. This might seem unnecessary in an office building that maintains constant levels of temperature and humidity, but there are several important factors to consider. One of these factors is whether the comfortable climate you feel during working hours is maintained around the clock. Some offices shut off the heat and air conditioning at night, and any computer equipment that is left running could easily overheat on a hot summer night in a sealed building. Excessively cold temperatures can also affect the performance of computer equipment.
Consider also that the equipment closets in which you plan to locate your servers and other components might not be climate controlled. If you plan to create a data center containing a large number of computers, routers, and other heat-producing devices, you will probably need an independent climate control system to keep the room cool enough. A source of clean, consistent electrical power is also important. A data center might require its own electrical circuits to support a lot of equipment in a single location. If this is the case, you might also want to consider adding surge protection and a backup power supply for the entire installation. This could include standard uninterruptible power supply (UPS) units, or even a backup generator with a failover switch.
Of course, not every network installation is located in a comfortable office building. You might find yourself having to plan for a network that is exposed to extreme environments, such as outdoors or industrial areas that expose the equipment to abnormal amounts of heat or cold, humidity, dust, electromagnetic interference, chemical vapors, and so forth. There are a variety of products available that enable computer networks to operate in these conditions, including waterproof keyboards, computers with air filtering systems that keep out dust and other contaminants, and networking technologies to suit a variety of conditions, such as fiber optic and wireless media.
If you will be performing an internal cable installation, you need to examine the places where the cables will run for possible obstructions and sources of interference. Copper-based cables are highly susceptible to electromagnetic fields caused by fluorescent light fixtures, electric motors, and other types of electrical equipment. Even seemingly benign factors like the locations of television sets, radios, and electric heaters can affect network performance. If you will be upgrading an existing network, you might find that these environmental factors are the cause of error messages or other problems the network is experiencing.
To ensure that your network is reliable, check to see that there are cable routes you can use that avoid these sources of interference. Otherwise, you might have to use cable with additional shielding, or even fiber optic cable, which is not affected by electromagnetic interference. You must also be aware of the fire and building codes in your area that govern the cabling process. You might need to purchase cables with a special plenum-rated sheath to run them through the building's ventilation spaces (which are called plenums).
Even if you are planning to outsource the cable installation, you should be aware of conditions that can affect its cost.
After you have determined the organization's networking needs, you can start to design the network and begin selecting the products you'll need to construct it. For a brand new network, you might be responsible for the computers themselves. You must be careful to select models that can fully support the networking tasks expected of them. If you will be networking existing computers, consider their age and whether you need to upgrade them by adding memory or disk space. It doesn't make sense to install a state-of-the-art, high-speed network to connect a bunch of old 486 computers. Depending on the users' needs and the budget, you may have to settle for a more modest network or purchase new computers.
Hardware compatibility is always a major issue when you're planning a network, but if you will be working with existing equipment—either an existing network or a group of stand-alone computers—compatibility becomes even more crucial. Many of the purchasing decisions you make for network equipment will be based on the protocols you choose to run, especially at the data-link layer.
Most Ethernet products work well together, even when they're made by different manufacturers, but you still want to be sure that the products you select all support the same type of Ethernet. For example, if you're expanding an existing network, you might want to use Fast Ethernet on the new computers and upgrade the existing computers from regular Ethernet at a later time. You might purchase dual-speed network interface cards (NICs) for the new computers, but leave the 10-Mbps NICs in the older ones. In such a case, you must purchase a dual-speed hub as well because, although a Fast Ethernet hub will support the new computers at 100 Mbps, it won't support the old computers at 10 Mbps. You can't connect a 10-Mbps hub to a 100-Mbps hub, so the old and the new computers won't be able to communicate with each other. A dual-speed hub with ports that can support either Fast Ethernet or regular Ethernet enables you to connect all of the computers to one network.
In addition to major compatibility issues like these, there are many small decisions you must make to ensure that all of the components of your network function together properly. For example, you must make sure that you purchase NICs that use the appropriate bus type for the computers and the correct connector for the network medium you've chosen. Also, make sure that there are NIC drivers available for the operating system you want to use. When you decide to use a particular grade of cable (such as Category 5), you must be sure that all of the cabling components that carry network signals, including connectors, wall plates, patch panels, and patch cables, are of the same grade.