Control, data, and management plane and operational design
What this lens evaluates
Whether each plane of the network is designed to scale, converge, fail gracefully, and be operated - and whether day-2 operations (monitoring, change, automation) are designed in, not bolted on. This is one of the heaviest domains in the blueprint.
The questions it forces
- Control plane: do the routing/switching protocols scale and converge within requirements? Are failure domains bounded?
- Data plane: how is traffic forwarded, queued, load-balanced, secured? Any MTU/encryption implications?
- Management plane: how is the network monitored, changed, and secured operationally?
- Automation/operability: what is the automation strategy, and is the design supportable?
Core considerations
- Control plane -> IGP selection - OSPF vs IS-IS vs EIGRP, BGP in enterprise design, Spanning Tree - types, tuning, and loop mitigation, Factors affecting convergence, Fast convergence techniques.
- Data plane -> Transport protocols overview - TCP, UDP, QUIC, Buffering, microbursts, and congestion signaling, Load balancing and hashing - ECMP and entropy, MTU, fragmentation, and PMTUD.
- Management plane & operations -> Network management - traditional vs model-driven, Infrastructure hardening - control, management, and data plane.
- Automation strategy -> Network automation fundamentals - why and the operating model, Telemetry-driven and closed-loop automation, and the AI counterpart AI-assisted operations - AIOps.
The AI/ML angle
AI fabrics stress this lens hard: the data plane must be lossless (Lossless fabric - PFC, ECN, DCQCN), the control plane is usually an eBGP CLOS (Connectivity models and SDN in AI fabrics), and operations rely on PTP-aligned telemetry + AIOps to find stragglers (PTP timing for AI fabrics, Latency and the straggler problem).
Architecture frameworks and KPIs
Designs sit inside business and enterprise architecture: business-architecture guiding principles and KPIs define what "good" means and how it is measured, while enterprise-architecture methods (e.g., TOGAF) and operational frameworks (FCAPS, ITIL, DevOps) govern how the network is run and evolved (Reference models and migration considerations). Tie operational design back to measurable KPIs so the management plane proves the design is meeting business outcomes - not merely that links are up.
How it shows up in a scenario
You will be asked to choose/justify protocols, tune convergence, and design operability. Reason plane by plane and treat supportability as a scored requirement, not an afterthought.
Tip: Scenario reflex
Read the supplied operational data before you decide. Utilization graphs, telemetry, and capacity figures in the brief exist to be used - a near-saturated link can eliminate an otherwise-attractive option outright (e.g., you cannot terminate new SD-WAN tunnels on an already-saturated DC Internet link). Check current load and headroom against the proposed change first.
Spaced repetition
What should you always check in the supplied data before recommending a change?
Current utilization / capacity / telemetry - a saturated link or no headroom can rule out an option outright (e.g., new tunnels on an already-saturated DC uplink).
What are the four "planes/aspects" this lens walks?
Control plane (protocols/convergence/scale), data plane (forwarding/QoS/LB/MTU/encryption), management plane (monitoring/change/security), and automation/operational design.
Why is operability scored in CCDE, not optional?
A design that can't be monitored, changed, or supported fails the business - day-2 operations and automation strategy are first-class design goals.
In AI fabrics this lens means [...].
In AI fabrics this lens means lossless data plane (PFC/ECN/DCQCN), eBGP-CLOS control plane, and PTP-aligned telemetry + AIOps operations.
Why connect operational design to KPIs?
KPIs (from business/enterprise architecture) define and measure "good," so the management plane should prove the design is meeting business outcomes - not just that links are up.
Sources
- Cisco CCDE v3.1 Unified Exam Topics (Domain 2).