Planning and designing a network
This chapter provides the needed background information to design a network for a school district. It begins with a discussion about computer networks. This provides a novice computer user with the background information required to read the rest of the guide. After the background on computer networks, a section lists the steps involved in developing and implementing a district network. Since many districts may feel overwhelmed by the amount of knowledge needed to design a district network, the following section discusses some of the different resources available to help a district as it is starting. The next section lists the steps involved in developing a networking plan for a district. Next, a section discusses the personnel needed by a district to implement and support the network. The chapter concludes with suggestions about how to train district staff on ways to take advantage of the opportunities a network provides.
2.1 What is a computer network
In the simplest terms, a computer network is a collection of computers that communicate with each other to allow the sharing of information and equipment. The information might be something as simple as a staff memo or something as complex as a student record database. Shared equipment might include anything from a disk drive and a printer to an entire building of computers. The advantage of a network is that a student could just as easily share his work with another student down the hall in another class as with a student halfway around the world. Without the network, the student could only share their work with those in their class.2.1.1 The Internet
The Internet, the largest network in existence, carries most of the information shared on a world-wide scale. It currently connects 20 to 40 million machines located throughout eighty percent of the world, as shown in Figure 2-1, and is growing at an astounding rate. International Data Corp. (IDC) estimates there were 56 million Internet users at the end of 1995, up from 38 million Internet users in 1994, an annual growth of nearly 50%. IDC estimates that another 200 million users will join the Internet by 1999. With this many users, the Internet has grown far beyond the expectations of its designers when it was created 27 years ago.
In 1969, a group of researchers created the precursor to the Internet as a testbed for research. They named it ARPANET after the Advanced Research Projects Agency (ARPA) which provided the funding for the project. Besides research, several universities, military bases, and government institutions used it to exchange electronic mail (email) and files. It remained under the control of ARPA until 1975, when the Department of Defense assumed control of ARPANET. The Department of Defense left it untouched until 1983, when they split it into ARPANET and MILNET. Today, the Internet has completely absorbed ARPANET while MILNET continues as the military segment of the Internet.
2.1.2 Wide Area Networks and Local Area Networks
The Internet is an example of a Wide Area Network (WAN), a network that transmits data over large distances. The other major type of network, called a Local Area Network (LAN), usually occupies only a single building or small campus. Two measurements compare the performance of these two types of networks, the maximum speed or bandwidth of the network, and the delay across the network. The bandwidth measures how much information can be sent across the network in a second. The delay measures the time from information transmission to its reception.
LANs typically cover smaller areas such as a single building or a small campus. LANs are usually faster than WANs, ranging in speed from 230 Kbps up to and beyond 1 Gbps (billion bits per second) as shown in Figure 2-2. They have very small delays of less than 10 milliseconds.
2.1.3 Packet switched and circuit switched networks
When sending messages from one machine to another, computer networks break the messages down into packets, the most basic unit of information used on a network. The type of network dictates how these packets get from an originating machine (the source) to a receiving machine (the destination). A packet switched network is a type of network in which each packet contains the source and destination of the packet as shown in Figure 2-3. Each packet is sent independently from the source to the destination. This is similar to how the post office sends letters. When several letters are sent to the same destination, each letter may not take the same truck or plane, but they all end up at the same destination.
Another type of network is called a circuit switched network. This operates in a manner similar to the phone system. When two people call each other on the phone, the phone company creates a connection when the first person dials and maintains the connection until either person hangs up the phone.
2.1.4 Communication protocols
The Internet uses a protocol called TCP/IP (Transmission Control Protocol/Internet Protocol) to allow communication between machines connected to it. A protocol is very similar to a human language, and just as many languages exist, many protocols also exist. Some allow specific pieces of hardware to communicate, while others allow communication of complex messages such as email. The protocol defines what messages can be sent, the structure of the messages, how to start and end communication, as well as many other elements needed for meaningful communication.
2.1.5 Services available
Even when computers communicate using a common protocol, they still need to decide what information to exchange. Services provide a mechanism for exchanging information between one machine to another. A service is a function provided by one computer that other computers can access. The term service grows from the concept of a client-server model, in which one machine, the client, asks another machine, the server, to perform a task for it. For example, a guidance counsellor could use client software installed on their computer to ask a student record server where a student currently is, and the client could then display the result for the counsellor (see Figure 2-5). By having all the clients access the same server for the student's records, only the server has to keep the student's information up to date. The rest of the machines can get that information from the server.
Getting a student's records is only one type of service, and many other services exist which a district-wide network can provide. These include file and print services, multimedia services, email, web access, Usenet newsgroups, an electronic phonebook, video conferencing, and administrative record keeping, as explained later in this section.
2.1.6 Network topologies
The topology, or physical layout, of the network provides another way of differentiating between networks. Different network topologies offer different advantages and disadvantages in cost, complexity, and robustness. The first two differences are self-explanatory and the robustness of a network is its ability to continue functioning even if damage occurs to part of the network. The three most often used topologies are the bus, the ring, and the star. In addition, districts can create and use combinations of the topologies as needed.
The next topology is the ring. It is very similar to the bus, except that the two ends of the network connect to form a complete ring as shown in Figure 2-7. The chief advantage over a bus is that if a break occurs in the ring, the machines will still be able to communicate by going to other way around the ring. Unfortunately, the difficulty and cost of bringing both ends of the network together and wiring a ring topology usually outweigh the advantages of using a ring topology.
The third basic topology is the star. In this layout, the network designer designates one or location as the center of the star. They place special equipment called a hub at this point and then run connections from each of the machines to the hub as shown in Figure 2-8. By connecting all the machines directly to the hub, a break in a single network cable will, at worst, cause one machine to become unable to communicate with the others. Unfortunately, failure of the hub would prevent any of the machines from communicating. Choosing good hub equipment can minimize the risk of this occurring. The disadvantage of using a star topology is the large amount of wiring that it needs during implementation. If the designer places the hub far from the machines it connects, the cost to run multiple lines to it instead of just one or two can be high. By choosing a central location for the hub, the designer minimizes the costs and the added robustness of a star will make up for the additional cost.
Along with the three basic topologies, a designer can create hybrid combinations. The most common is the hierarchal star. In this layout, the designer runs only one line from the hub to a classroom where they place a second hub. They then connect all of the machines in the classroom to the second hub. This reduces the amount of wire needed to create a true star, although it increases the risk of a single network failure affecting many machines. In another common layout, the designer runs a single connection from the hub to the classroom. They then connect all the machines in the classroom together in a bus, and connect one end of the bus to the hub connection.
2.2 The networking process
The complexity of designing and implementing a district network can even overwhelm a person with a strong computer background. By having a good understanding of the entire process and developing a plan for the entire process, a person supervising the design can keep things manageable. Keeping a view of the larger picture will also allow the designer to predict pitfalls before they occur and will allow better flexibility in solving problems as they inevitably will occur.
While the district is reviewing and allocating resources, spend the time looking for outside help. With the help of this outside expertise, the next step, developing a networking plan, will be much easier. During this step, create the actual network design for each building as well as for the district. Upon completion of the plans, a district can implement them. This step includes activities such as wiring buildings and connecting buildings.
2.3 Resources available for help
Networking a school can be a large and complex project. Sometimes this necessitates the need for outside help. The help can come in many forms, including money, labor, and knowledge. It can make the difference between getting the network to work or not work. Districts can find help found locally, from statewide or nationwide organizations, and online.2.3.1 Locally
The parents of students in the district and the students themselves are a local resource. Parents with computer experience are often willing to share their knowledge and experience. Parents and can also provide money through activities such as fund-raising. Shown the benefits of new technology to their children, parents are often willing to help. Technology companies in the area can provide both equipment and services for the district in return for publicity for their donations. 2.3.2 Statewide and nationwide
Along with help from local resources, the state or federal government can also provide help. Many state boards of education plan to implement statewide networks and to provide access to this network for districts. For example, the Illinois State Board of Education (ISBE) currently plans to create a statewide network connecting all the regional offices around the state. Their goal offers Internet access to districts through the regional offices. They have also established Technology Learning Centers at several of the regional offices. Experienced network staff work at the centers and can provide technical support to a district. Although these offices are only for Illinois schools, similar initiatives are planned by other states.2.3.3 Online
Besides getting help and money from people, companies, and government institutions, districts can look at what other school districts have done. Districts can look at this information online, as most districts want to announce their presence on the Internet as soon as they connect.
2.4 Developing a networking plan
After understanding the concept of a computer network and seeking outside help to aid in the design process, it is time to develop the actual network plans for the district and each building. The plans should focus on the design of the network, not the implementation specific details such as equipment brands and models. The designer will add those specifics later while implementing the plan. Figure 2-10 shows a flowchart of the process of developing the networking plan.
During the development of the plan, the designer should consider ideas such as future upgradability. The infrastructure installed should be suitable for the advances of technology over the next 10 to 20 years. Although this will cost additional money, it is better to provide the necessary infrastructure the first time rather than having to replace it in a few years. Table 2-2 shows a list of recommendations for a network plan that will provide for future growth.
2.4.1 What to buildings to connect
The first thing that needs to decide before beginning the network plan is which buildings are going to connect. The district will need to decide if it wants to connect all the schools, administrative offices, and other district owned or run locations. One approach is to connect a part of all the buildings to the district WAN initially, with the rest of each building being networked later. This is the method that Champaign chose to implement because they were installing a new administrative system that required that a connection be established with guidance counsellors at each school. They later went back and finished networking the remainder of each school. Another approach is the one being used by Urbana. Under this method, they completely networked each building and connected it to the district WAN before work beginning work on the next building. Table 2-3 lists the pros and cons of each of these methods.
2.4.2 What rooms to connect
Once the buildings to connect have been determined, the next step is to decide which rooms to connect. Again, depending on the needs of the district, it can either implement an all-at-once or a phased design. In either case, the district needs to decide which rooms to eventually connect.2.4.3 Where to place the network hub
After deciding which rooms to connect, the designer needs to determine where to place the network hub. When determining the location, there are a number of factors to consider. The first is that the location should be centralized. This minimizes the amount of wiring needed to connect the entire building as well as it avoids exceeding any of the wiring length restrictions discussed later in the Chapter 3. The location should be isolated from the general building staff and students to prevent accidental or deliberate tampering. If possible, it should also be environmentally controlled with continuously running ventilation, preferably with an air conditioning system. It should not contain any water sources such as sinks, steam pipes, or water pipes. Finally, it must be large enough to hold all the equipment that will be installed as well as any future equipment that a district will install in the foreseeable future.
If a single location can not serve the entire school, as happens in large buildings, then a district can use multiple locations. Each hub should be located centrally in each part of the building, and the hubs should connect to create a single network. When connecting them together, remember that there will be a significantly larger amount of traffic between the hubs than along the normal wiring to each classroom. For this reason, or if the hubs are located further apart than allowed by normal wiring restrictions, consider connecting them with fiber optic cable instead of normal wiring.
2.4.4 What technology to use
Although networking equipment has not been discussed yet, considerations need to be made to ensure that a district can implement the network plans using available equipment and technology. A district should select the best technology it can currently afford. It should begin with cheaper and slower equipment, while still designing a network infrastructure that it can easily upgrade to faster and more expensive equipment in the future.2.4.5 Services to provide
As already discussed, there are many services available to a district. A district will need to base these either at some or all the schools, or at a single central location. With some of the services such as an administrative system and a district phonebook, a district can easily decide they need to be centralized. With others such as file, print, and multimedia servers, a district can place them at each school. With a third group including email, web, and news, the district can not easily make a decision. From one perspective, these services are going to be used at each individual school and this would make it appropriate to place them at each school. However, what happens when a student changes schools? This would require that all of a students electronic information be moved as well. A student's email address and web URL would also have to change since they derive from the student's school. By using a centralized location for these services, this information would not have to change or be moved. The student's email address would derives from the district, not a school. Although this would seem to be a better solution, from a technical perspective it is inferior. The individual buildings will likely connect with a slow speed WAN. When trying to access information on a central server, a student could wait much longer than if the server had been placed locally at the school. Costwise, all of the equipment will need to be duplicated at each school, increasing costs substantially.2.4.6 Internet connectivity
Providing Internet connectivity should be one of the primary design goals of the network plan. This includes getting a connection to the Internet, registering and maintaining Internet information with the correct Internet authorities, and installing and configuring the correct software on each machine needed to access the Internet.
where <DISTRICT NAME> is a name chosen by the district, and <STATE> is the state in which the district resides. For example, Champaign and Urbana share a domain called CMI.K12.IL.US, and Mahomet-Seymour has a domain named MS.K12.IL.US. Besides registering the domain name, the district will need to get a range of IP addresses that will allow the machines in the district to communicate with others on the Internet. The InterNIC or the ISP can provide this range of numbers. The choice will depend on the ISP, but be sure that if the district does need its own numbers that it applies for them. In either case, it is recommended that a district request a single Class C address range for each high school, one range for every two middle schools, and one range for every four elementary schools. Each Class C address provides IP addresses for about 250 machines, so this will provide high schools with 250 addresses, middle schools with 125 addresses, and elementary schools with 60 addresses. These numbers should offer many more addresses than machines currently at each school and will provide room for the school to grow. More information about domain names and IP addresses is available on the InterNIC web site at http://www.internic.net/.
2.4.7 Security
When students begin using the network, security will become an issue. Students may begin trying to access forbidden information such as student records. Providing security to district machines while still providing students access to the Internet is, unfortunately, not possible. Increasing one forces a decrease in the other. At one extreme is a completely open network in which students can do anything they want and have the freedom to explore the entire Internet. At the other extreme is a completely closed network in which students can only access sites that the network administrator has allowed. The first places the district's machines at risk, while the second can inhibit student creativity.2.5 Personnel needed
The personnel needed by a district, to support a network, fall into the four categories listed in Table 2-4. This first category contains the people who will design. The second are the ones who will install the network. The third are the staff needed to support the network's day to day operations once it is running. The last group are the ones who will train the district staff on how to use the network. Although the same people can be performs all four jobs, having multiple people involved will allow each one to specialize on their specific job requirements. Urbana has two full time staff members involved in the networking process, one whose job focuses on the technical details of supporting the district network and the other whose job focuses on staff training. Of course, for major problems they work together, but the division of labor allows each one to concentrate on their specific area.
2.5.1 Design personnel
The design phase usually has limited personnel requirements. The work does not require a large time investment in each design, but must to be done correctly for the network to function. Having outside help and several people to share ideas with will help produce the most cost effective and technically superior designs. These people usually hold positions such as district technical coordinator. They need to be able to combine the network requirements of the district with the realities of what a network can accomplish.2.5.2 Installation personnel
During installation, people will need to run wiring throughout the school and to perform other manual labor tasks. The staff needs during installation are usually short term. Many people will be needed while networking each building, but once completed, the personnel can be reassigned to other jobs. There are many choices where the staff can come from. Urbana chose to use district electricians to install the equipment, reducing their cost by using existing staff. Mahomet chose to do the entire installation with parents and other volunteers working under the guidance of the network designers. Champaign, because of the large number of buildings to network, chose to hire two new full-time staff members whose jobs were to install the network at each of the buildings. Charleston chose to hire outside contractors because they did not feel they had the experience to install the network themselves. 2.5.3 Support personnel
After installing the network and verifying that it functions properly, someone will need to maintain it. These duties include troubleshooting problems, repairing broken hardware, reconfiguring software, backing up important data, updating software as new releases are available, and planning upgrades to the equipment as technology improves. This is often too much for a single person or even a couple of people. If one person has to support multiple buildings, they end up wasting a lot of time traveling between the locations. Instead, a district can take an approach like Urbana's. In Urbana, one existing staff member at each building trained in the basics of computer networks. They act as the building network administrator. If a problem arises beyond their abilities, they act as a liaison to the district support staff for additional help. This method places a person at each building, alleviating much of the wasted travel time. It also allows the building staff to become more familiar with the building than would be possible using a central support person who supports many buildings.2.5.4 Training personnel
Along with supporting the network after installation, someone needs to train the district staff on how to use the equipment. Often it is best to take a pyramid approach to training. In this approach, a group of interested staff members are taught about networking and computer basics. Then this group is sent out to each school to teach the staff at the buildings. This allows for a rapid dissemination of information, although it can result in unqualified people training other people. By carefully selecting and training the first group of trainees, a district can avoid this problem.2.6 Staff Training
As has already been mentioned, training the district staff in the use of the new computer equipment and network is essential to the network being used once installed. The training should begin when some of the network is up and running, especially if a district is going to train a group who will later train the rest of the district. Figure 2-11 shows the process of developing a training program.
2.6.1 What to teach
What classes should a district teach? There will be staff ranging in experience from never having used a computer all the way to ones who are already accessing and developing for the Internet from their home. To handle the range of trainees, a district should develop a program which teaches everything from the basics all the way to advanced computer curriculum development classes. If possible, share the burden of teaching with other surrounding districts. Urbana, Champaign, Mahomet, and Tolono have pooled their resources to offer a variety of classes over the summer, not just to their staff but also to the surrounding community. The list of classes in Table 2-5 shows that by pooling their resources they are able to offer classes to just about any level of experience. Besides offering as many classes as possible, a district should also revise the class list as new technology and opportunities arise. For example, if the district installs a new multimedia lab, it would be wise to offer a training course in its use as soon as possible.
2.6.2 Who to teach
Although many staff members will want to take these classes, others will have no time or interest in taking them. A district will need a policy in place stating which teachers are required to take them and which ones are not as well as why. The policy should also indicate the required classes. If a district feels that basic computer competency is enough, it may only require the introductory classes. The real minimum should be at least a level where a staff member can take advantage of the new network. This will require classes in computer, email, and web use. 2.6.3 Logistics of training
Once a district determines the classes to be taught and the class requirements of the staff, it needs to plans the specifics of training. Classes can only be taught when school is not in session unless substitutes are available while the staff members are in training. This leaves staff development days, school improvement days, and summer break as viable times to perform the training. Some of the buildings may be locked over the summer, may not have air conditioning, or may already have other commitments. A district needs to sort all of these conflicts so that they can present the staff with a cohesive class schedule that the staff can select from.
Return to the Table of Contents
This file last updated on 05/09/96 at 13:43:48.
© Copyright by Bradley H. Lamont, 1996