Once the customer requirement have been finalized, all goals and constraints defined, and the state of the current network assessed, it’s time to begin designing the new or upgraded network. Earlier in this chapter the two major approaches to network design were discussed, and you should already be aware that the top-down approach is preferable in almost all circumstances. Using this method, a network is designed with the specific application and service requirements of the customer, as well as a very high-level concept of the network infrastructure that will be needed. In particular, the top-down approach should begin by considering the following general requirements:
Functionality. The new design must meet all of the functional requirements for different applications and services as defined by the customer, and augmented by your analysis. All of the functional requirements of the network, such as the need to support VoIP or IP Multicasting, should be defined and well understood at this point.
Performance. The performance requirements for the new or upgraded network must also be defined and understood as part of the design process. All of the proposed applications and services must be understood with respect to the impact that they will have both individually and together on the new network. Other examples of factors that will influence the performance of a network include the number of users that the network needs to support, the speed of the network, and so forth.
Scalability. Although current needs usually have the biggest impact on network design, it is imperative that planned future expansion also be considered. For example, a company may be planning to implement new applications or services in the next 12 months, or be planning to merge with another organization. The design must factor in these elements, including how the new design will allow the network to scale to meet these needs as necessary.
Availability. The availability requirements of a network also have a tremendous impact on any network design project. If a customer’s network or certain applications and services are considered to be mission-critical in nature, the new design must take factors like redundancy into consideration. This may ultimately involve connections to multiple ISPs, the need for redundant switching paths, and so forth. At this stage in the process, a designer is not usually worried about the specific technologies or methods to be used to ensure this availability, but instead concentrates on defining where redundancy needs to be considered.
Security. The security of a new network is a critical component of any design. In the top-down approach, the security needs of an organization should be considered at the beginning of the design process rather than as an afterthought. Again, elements like specific firewalls or intrusion detection systems are not the focus at this point. Instead, the network designer should be focused on the specific network applications and services that need to be secured, and how these requirements will impact the rest of the design process.
Budget. As mentioned in the business constraints section earlier, any network design project needs to consider any budgetary limitations that exist. The designer must be aware of and consider these limitations at each stage of designing an appropriate solution, based on the importance that the customer has associated with various planned applications and services. Ultimately, the budget associated with the project will impact everything from the equipment chosen through to project staffing.
After considering the general network requirements listed above, a network designer using the top-down approach would continue by organizing the network into what are referred to as “modules”. A module is a distinct portion of a network that is responsible for a specific and particular role. For example, the WAN module would be concerned with the equipment and links that would make communication with other locations possible. The network management module would be focused on the applications and equipment required to manage and monitor a network, such as a network management system (NMS) like CiscoWorks.
The model used by Cisco to segment a network into distinct modules is known as the Enterprise Composite Model. This model, along with the functional areas and modules it is comprised of, will be looked at in more detail in future articles.