The purpose of this case study is to develop a network for the government agency involved in providing jobs to the unemployed. It is headquartered in two buildings where there are offices of the various departments. The agency is being led by a director, an assistant director, and the executive secretary to the director. The purpose is to provide employment opportunities to the youth and adults in the community. Accordingly, it matches the job profiles that it receives from the unemployed with the vacancies that are available. However, most of the times it is determined that the unemployed are not adequately skilled to take on those jobs. Accordingly, there is a need to train them in essential skills required. Resultantly, it also then provides them with fundamental training in basic office skills, such as word processing software, spreadsheet software, elementary website design and maintenance, and basic IT support. Then, these individuals are ready to take on the various jobs that require these skills.The agency has five departments in all that provide the various services required by this personnel. These include accounting, testing, intake, training, and job market analysis. There is a number of people that provides various services in these departments. Each department is led by a department head. There are also numerous training facilities available at different locations in the two buildings. These have been provided with personal computers for the use of trainees. Additionally, the staff members also use personal office computers. However, the problem is that they are not connected with each other or with the Internet. Accordingly, the agency has sought the consultants to develop an IT network for them. This network is expected to provide the essential IT networking services to the organization that would include the provision of a LAN (Local Area Network), connectivity with the Internet, and provision of email service within the organization. It has been observed that there are currently only 27 workstations in the agency use. Accordingly, the agency has not asked for any scaling up of the current infrastructure. They only want to include the network for these current workstations. Consequently, no future availability of the infrastructure is required (Neat, 2004). The proposed networking infrastructure only has to connect the workstations currently in use.
A network topology is a map of the internetwork that shows the segments, the interconnection points, and the user communities (Naimzada, Stefani, & Torriero, 2009). It is a blueprint of the network geometry which does not mirror the actual physical layout of the network. It provides the ability to achieve technical implementation and it is a high-level blueprint that shows the architecture of the network and not the actual details of its implementation.
A network topology design is the first phase of the top-down network design methodology. When a network topology is identified before selecting the physical products and network components it provides a way to achieve adaptability and scalability for the customer. The process identifies networks and interconnection points, types of internetworking devices that will be used, and the scale and scope of the networks. It does not comment on the actual devices that will be bought.
Because the network implementation for this agency will involve implementing the network in two separate buildings with possibilities for future connectivity with some other remote facility, it is essential to include a WAN design in the network architecture. Although a LAN can also be implemented in the agency without the need for a WAN, it is required for the scalability of the internetworking architecture. Otherwise, it would not be possible to extend the LAN to any other facility that is added to the organization. The requirement for including a WAN is understood even in a small organization like this one. Furthermore, the provision of the Internet facility to multiple nodes also requires implementing a WAN. Accordingly, in such a condition the only viable network topology is a hierarchical network design.
The technical and business needs of the organization will only be met once it has a network topology which has multiple interrelated components. Likewise, the topology is developed through 'divide and conquer' methodology which breaks down the network into its constituent layers. The hierarchical network design provides the network in discrete layers. Each layer has its own functions which can be provided by selecting the most appropriate systems and devices. Here, we will not need high-speed WAN routers as we do not have a WAN backbone currently connecting multiple locations. The two buildings are in a single location. Therefore, we will need medium-speed WAN routers to connect the two buildings. Then, the servers and user devices in each building will be connetorough the LAN provided by switches.
The hierarchical network topology consists of three layers. These are the core, distribution, and access. The core layer consists of high-end routers and switches which have been optimized for performance and availability. On the other hand, the distribution layer provides infrastructure that implements the Information Systems (IS) policies, such as those for security. It also contains routers and switches. Finally, the access layer connects the users with low-end routers and switches and wireless access points.
If we do not use this network hierarchical topology then the network will grow in a haphazard unstructured manner. The unplanned networks are also referred to as fur-ball networks. A fur-ball network has many CPU adjacencies which clog the network. This happens when the devices start communicating with many other devices that is very burdensome for the CPU. As an example, if we do not include the core layer here containing the medium-speed WANs, then the network architecture will be that of a flat-shaped WAN. This will be an entirely switched network providing the LAN with no core and distribution layers. Here, a broadcast packet will ask for processing time from each device in the network, thereby slowing down the entire network. It will slow down all the devices in the network, including the switches, nodes, and the servers.
This network clog is in addition to the CPU workload that is required by it. Then, the CPUs slow down as they have to process each router request in a flat WAN topology which does not have a core and distribution layers division. In this way, the hierarchical network topology prevents such situations by implementing a modular design where all of the devices do not unnecessarily communicate with each other. However, they are still connected with every other device on the network.
This model also minimizes the cost as there is no need to implement specialized devices for the single or two layers in the architecture. Then, only general equipment and devices are included in each of the three layers. It also helps in estimating the capacity and subsequent accurate planning for each of the three layers in the network, thereby saving on the bandwidth. Finally, the network management costs are minimized as the network management for each of the three layers is separate in the modular network architecture.
Then, each and every design element is also simple and easy to understand. This simplicity is not only required for quick and convenient implementation, but also simplifies the work of network management and administration personnel. It is also easy to test a network quickly which is modular in nature with discrete layers. The transition points become conspicuous and it becomes easy to find faults in the network to identify the failure points.
It is also very easy to implement changes in such a network topology. When a device or a component of the network needs to be changed, the cost is limited to replacing only a small subset. In large flat or meshed architectures, the changes impact a large number of systems and components. Just replacing a single device becomes cumbersome as the device is interconnected to a large number of the other ones (Oppenheimer, 2004). Consequently, we will be implementing a hierarchical network topology for this network.