Introduction to Top-Down Design Approach
Date: Mar 14, 2011
For all those people already bored by the title of this article, I will attempt to make this concise enough to be interesting and help in everyday network designs. Network design, to many people, is something that sounds interesting but can also be construed into a boring, analytical task that is less interesting to the “typical” network engineer who wants to get their hands dirty. Some of this is true; the design phase of most projects does require a level of analysis and analytical tasks that require an understanding of not only the technical aspects of the networks but also the business factors that affect the network.
This article takes a look at the concept of Top-Down network design. This is one common method used to design networks where the focus is put on the needs of the business first. Top-Down design is not a process that is limited to network design; it is a concept that can be used for many other applications. Top-Down network design is simply the analysis of the needs of companies from a high level. Instead of starting a design by choosing the equipment, the design starts by taking the company requirements and needs analysis information and using it to derive a design based on high-level functional requirements.
The following sections will offer a brief review of the different phases that are used with Top-Down network design; these include:
- Analyzing Business Goals and Constraints
- Analyzing Technical Goals and Tradeoffs
- Characterizing the Existing Internetwork
- Characterizing Network Traffic
Analyzing business goals and constraints
One of the most important parts of designing a network includes a careful review of the existing business goals and constraints. Often when these are not taken into account early in a project, unneeded changes must occur mid-project that can greatly affect the schedule and budget of a project.
When reviewing the business’s goals, it is important that as much information be gathered from as many sources as possible in order to have a complete picture of the business from multiple angles. For example, if the view is only from above, many things that happen in the trenches that are important can be missed, and vice versa.
It is at this part of the project that the scope of the project must also be calculated. How the project is being handled and whether it is supported by all departments within the company that will be affected are important questions. If not, it is important that any interdepartmental issues be addressed before the project starts; this will ensure a smoother project transition. These issues are also addressed below as business constraints.
It is also important during this phase of the project to get a detailed understanding of the current network applications, as well as any additional network applications that will be sought in the future. This information is vital as part of the design process from a business and technical perspective.
Another equally important part of this phase is analyzing business constraints. There are different aspects of network design that must be carefully completed; these include a careful accounting of the politics within the company. Many projects are hindered or halted by different political factions within a company. It is vital that each of these factions be happy when a project affects their areas of operation. Budget, staffing, and scheduling issues must also be carefully addressed in order to ensure that the project has the resources available to successfully complete the project.
Analyzing technical goals and tradeoffs
The next step to be addressed when following a Top-Down approach is to analyze the technical goals and tradeoffs.
The scalability of the existing network must be determined in order to calculate the required steps that must be followed to perform the additional design work. It is important that any design being put together have some level of scalability and be able to expand to the future requirements of the company. Along with scalability is the importance of a design’s modularity. When the network is put together using a modular approach, implementation, operations, and troubleshooting can be considerably easier. Cisco offers its modular solution inside their SAFE offering.
The availability requirements of the proposed network must also be addressed. For example, is it important for the network to have a near perfect up time? Along with figuring in the availability requirements of the proposed network is designing a disaster recovery plan; it is during those unexpected times when a plan can make or break a company. Another part that goes hand in hand with the availability requirements is the calculation of the performance requirements of the proposed network.
There are a number of performance calculations that must be completed in order to correctly propose the best solution, these include:
- Bandwidth requirements
- Delay/Response time requirements
- Jitter requirements
- Packet Loss requirements (goes with availability)
In the modern technical world, the security requirements of the proposed network are also of the utmost importance. The breach of security of any company can be catastrophic for a number of different reasons; it is best that all the security requirements of a new network be addressed early in order to account for the different potential security scenarios. As the steps required for the security of a new design can be complex, a thorough review will not be completed in this article.
The manageability requirements of the proposed network can also be easily overlooked. While it may be obvious that the elements within a network be manageable, it is also important to consider how the management of these devices will work. For example, what would happen if a failure occurred within the network? Another aspect to consider is how the manageability options of the current and future equipment could be used to improve other aspects of the network including performance, troubleshooting responsiveness, and day-to-day operations.
Characterizing the existing internetwork
The third step that that is used when following the Top-Down approach is the characterization of the existing internetwork. The information gathered from this step can greatly affect how a design is put together. For example, is the current equipment deployed in the network able to be used in the proposed design or does the proposed design require a complete network device change?
Different pieces of information to be gathered include:
- Existing topology map
- Existing addressing and naming
- Existing wiring
The existing topology map (if it exists) is a nice thing to have when designing an addition to an existing network. From this map, the designer can visualize the changes (if any) that are required to the existing network in order to have it integrate with the new proposed design. Sometimes this means that the existing network is well designed and requires few network changes, and other times it means that the existing network must be bypassed in order to replace it with a better designed network.
The existing addressing and naming of a network can greatly affect how a new proposed design is implemented. It is often required that the existing network address and naming structure be altered in order to improve scalability and modularity.
The advantage of a good existing wiring base is that it can potentially be re-tasked to run newer networking standards. For example, an investment two years ago in a Category 6 UTP wiring infrastructure could greatly reduce the cost of a network upgrade today.
Another issue that has become more important with the deployment of wireless networks is to find out any architectural or environmental constraints that exist within the buildings housing the network. For example, is there an abnormal amount of interference that could affect the way a wireless network can be deployed?
The health of the existing network is also a vital piece of information that is nice to have when designing a network. If the currently used network has certain health issues, these should be found so that they can be addressed as part of the new design. A good way to figure this out is through comparison of existing network baselines and/or performance metrics performed on the existing network. Network metrics should include current availability, bandwidth use, and delay, to name a few.
Characterizing network traffic
When designing a network that is replacing or building upon an existing network, it is important to characterize the current network traffic. This information can then be used to better customize the design of the new network.
A piece of information that is nice to have when designing a network is a network flows analysis. This includes a report that includes the current major network traffic sources and how their traffic typically flows across the network. A report and/or map of the existing traffic flows is also nice to have, as it provides a way to visually analyze the parts of the network with the highest load.
On modern networks that have or plan to have deployed voice and/or video solutions, it is nice to obtain a report of the current QoS parameters or future QoS requirements across the network. With this there would also be an analysis of the traffic types that are most commonly used on the network. As QoS is a vital part of any voice or video deployment, the assessment of these parameters is crucial in a new network design.
Summary
While network design for many engineers and consultants can be a very boring subject, it is also a main part that contributes to future network operations. Without a well thought out design, a network is bound to have future problems and be much more costly than is required. Putting a little more time and resources into a good network design not only makes the operations of the network easier but also saves money and time on future network troubleshooting and expansions.
Using a Top-Down network design approach allows the network to be more customized to the needs of the specific company. Without this customization, a network has the ability to become more of a hindrance then a help.
Reference
Oppenheimer, P. Top-Down Network Design (Third Edition). Indianapolis, Indiana: Cisco Press, August 2011.