Stakeholders, business drivers and the CCDE way of thinking
What you will learn
- How CCDE differs from a configuration exam: it tests judgement and defence of a choice.
- How to read a customer case and separate business drivers, requirements, constraints and assumptions.
- The four constraint families: technical, operational, application and business.
- How to list candidate designs, build a trade-off table and defend one choice, including the words "it depends on".
- How to prepare for pushback and learn from a wrong choice.
Prerequisites: CCNP-level knowledge of routing, switching and WAN technologies at a conceptual level (OSPF, BGP, VXLAN, SD-WAN). There are almost no commands in this module. Dual-stack (IPv4 and IPv6) is assumed throughout.
Think of two people looking at a house. The plumber asks which pipe to fit. The architect asks who will live there, what the budget is, what the land allows and what happens when the family grows. CCDE wants the architect. A design is "good" only relative to a business, so the first skill is reading the business.
Case file: Marigold Retail Group
Document 1, business brief (from the CFO and CIO). Marigold runs 210 stores and an online shop that already earns 28 percent of revenue. The board wants online revenue doubled in three years, WAN run cost cut by 20 percent, and self-checkout in every large store. Festival weeks such as Diwali bring about four times normal traffic and a large share of annual profit.
Document 2, current state. One MPLS carrier, one circuit per store, two data centres in active/standby, internet access only through the data centres, IPv6 only on the online front end. Six network engineers manage everything by hand. The MPLS contract has 14 months left and an early-exit penalty. The point-of-sale vendor supports only certain latency and fixed server addresses.
Document 3, operations note. Change freeze from 1 October to 15 December. No store may be offline during trading hours.
Step 1: extract requirements and constraints
A stakeholder is anyone whose goals or risks the design touches. A business driver is the reason the project exists. A requirement is something the design must achieve; it can be prioritised and sometimes negotiated. A constraint is a given that limits the choices; it cannot be designed away. An assumption is something you believe but have not confirmed, and it is a risk until confirmed.
| Statement from the case | Type | Family |
|---|---|---|
| Double online revenue, so the shop must cope with four times peak | Requirement | Business and technical |
| Cut WAN run cost by 20 percent in three years | Requirement | Business |
| Stores keep trading during migration | Requirement | Business |
| MPLS contract has 14 months left | Constraint | Business (contract) |
| Change freeze 1 October to 15 December | Constraint | Operational |
| Six engineers, mostly CLI skills | Constraint | Operational |
| POS vendor supports only certain latency and fixed addresses | Constraint | Application |
| Broadband in small towns is good enough | Assumption | Technical, to be tested |
Common mistake. Treating a preference as a constraint ("we only buy from vendor X") without asking who decided it and whether it can move. In the exam, a constraint the case states is fixed; a preference you invent is not.
Step 2: candidate designs and the trade-off table
| Option | Optimises | Sacrifices | Main risk |
|---|---|---|---|
| A. Keep MPLS, buy more bandwidth | Low technical and migration risk | Cost target, internet breakout, agility | Cost rises with the growth plan |
| B. Hybrid: an SD-WAN overlay over two transports in every store: MPLS plus broadband in large stores; broadband plus 4G/5G in small stores once their MPLS term ends | Cost, capacity at peak, agility, staged exit | Simplicity during the transition | Skills gap and two transports to run |
| C. Internet-only SD-WAN, cancel MPLS | Lowest steady-state cost | Availability guarantees, penalty paid | Weak broadband in small towns |
| D. Managed service from a provider | Skills gap, speed | Control, long contract | Lock-in |
Step 3: the chosen design and its defence
Choose B, with a roadmap that moves toward C store by store as evidence arrives. Defence: the contract makes an immediate exit expensive (constraint), so every store keeps its MPLS circuit until the term ends and gains broadband now; after that, large stores keep MPLS plus broadband and small stores move to broadband plus a 4G/5G path, so every store, large or small, always has two independent transports and meets the "no store offline" requirement; cheap broadband absorbs festival peaks; a controller-based overlay lowers the manual effort of six engineers; migration can be phased around the freeze. It depends on the size of the exit penalty and the measured broadband quality. If the penalty were small and the quality proven, C would win. A good answer states that dependency.
Step 4: what the customer will push back on
- CFO: "Why pay for two transports?" Answer: paying for MPLS and broadband together is transitional; after the MPLS term ends, the second path in each store is a cheap one, and it is the resilience the business asked for.
- CISO: "Local internet breakout widens the attack surface." Answer: add branch or cloud-delivered security to the design and cost.
- Operations: "We lack the skills." Answer: training and a pilot are part of the plan; if skills cannot be built, option D becomes more attractive.
Step 5: post-mortem of a wrong choice
A competitor picked option C because it had the lowest cost on one slide. Availability was never written down as a number. Small-town stores lost their only path during a festival week, trading stopped and the saving vanished in one weekend. The root cause was a missing requirement, not a bad router.
Every CCDE answer travels this path. Skip a step and the design defends nothing.
Exam trap. Do not name a technology before you have named the requirement it serves. A correct technology with no link to a stated requirement scores low.
The slide that hid a constraint
A bidder offered Marigold a fixed-price internet-only WAN. Reading the contract, the architect noticed the 14-month MPLS commitment and the freeze window, so the real timeline and cost were very different from the slide. The final plan overlapped both networks for a year.
Lesson: constraints often hide in contracts and calendars, not in diagrams.
"Where do you start when you get a design case?"
Strong answer: with the business drivers and stakeholders, then an explicit list of requirements, constraints and assumptions. Only then candidate designs, a trade-off table, a choice with a defence and a note of what would change the choice.
Key takeaways
- CCDE rewards judgement, so name drivers before technology.
- Requirements can be prioritised; constraints are given; assumptions must be tested.
- Constraints come in technical, operational, application and business families.
- Always compare at least three options and say what each sacrifices.
- State "it depends on" and what the answer depends on.