In this lesson we’ll cover Network Design Methodologies : The top-down approach and design life-cycles
When you’re finding yourself in the position of being responsible for designing either a change to the network or a brand new network or even a total redesign to an existing network, it helps to have a repeatable approach to the situation. That way when your boss says ‘we need a total redesign because we’re implementing VOIP!’ you’re not just like ‘uhhhh where do I start..’, you’ll know exactly where to start and what questions you should ask to start your design process.
So, Top-down vs. bottom-up; what do they mean here. I like to think of this in terms of the OSI model. So you’ve got your application layer at the top and going down to your physical layer at the bottom. With Bottom-UP, you’re shooting from the hip. In that VOIP scenario, some people are like ‘soo I know I’ll need some PoE switches, and a new router for voice traffic..’ and they’ll go ahead and start buying up all this hardware. They’re usually relying on past experience so they know the general idea of what’s needed, and they end up over-provisioning… by a lot, to make sure everything actually works since you’re not taking the extra time needed to go through the top-down process. This does allow for usually a much faster design time, but it’s not taking into account your future needs and goals as a company so a couple years, heck even a handful of months down the road you might be stuck doing a whole redesign, again, because the current one you just implemented can’t handle the growth.
With Top-down, we’re going the other way. Here we start with the application layer and work our way down, so we’re starting with the application’s requirements and what we’ll actually use and then work our way into finding out what hardware is most appropriate. Some questions you might ask yourself to answer are how much bandwidth is needed at peak usage, and what’s our expected growth. Sure, this requires a lot more work and usually much more time to complete, but you’ll get it ALL back, trust me! Even if just in the time you’ll save troubleshooting issues later because you took the time to plan out your design, and end up with a design that meets the company’s needs, is scalable and more future proof, and this next one is what you can sell the company on when they gasp that it’ll take you 3 weeks to make the design, is that overall it has a reduced total cost of ownership due to being properly scoped and requires minimal changes to scale with the company. I know, we’ve all been there and tried to be the hero. How do you become a hero as a network engineer? Come in under budget and finished early. Sure you might get something that works great for a month, or even a couple months, but eventually you’ll be kicking yourself that you didn’t put the time in to design the network properly and end up having more downtime that you even needed previously to correct the situation. So, top-down, not bottom-up 😊 though you should be aware of both because Cisco always loves to ask a question to make sure you understand the difference.
Now Cisco has developed 2 different design life-cycles. First, was puh-puh-dee-oh 😊 , an acronym typically spelled rather than said. PPDIOO formalizes the steps involved with designing and operating and eventually optimizing the network. This cycle consists of:
- Prepare – Determine organizational goals and constraints, create High Level Design (HLD)
- Plan – Audit current network and perform gap analysis. Determine technical requirements
- Design – BoM and LLD are created using information gathered in prepare and plan phases
- Implement – Prototypes and pilots are used. NIP is created to facilitate implementation
- Operate – The longest phase. Includes day-to-day operations, monitoring and troubleshooting
- Optimize – Proactive management to resolve problems before they become problems