Points: Tier 1 gate (pass/fail) + Tier 2 scored
Time estimate: ~16 hours (Weeks 13-14 plus independent sessions throughout)
Position: Lab 13 in the lab index; terminal NET-301 deliverable
Mission
The capstone integrates the four scales of NET-301 into one end-to-end exercise. A student who completes the capstone has designed and deployed a multi-site enterprise network at carrier and datacenter scale, reverse-engineered an unknown protocol observed on its traffic, and deployed NSM coverage that detects it. This is the portfolio artifact that demonstrates Belt-5 networking competence.
The Canonical Scenario
The instructor seeds the lab environment with one unknown-protocol stream on the student's network traffic. The student does not know in advance what the protocol is; they must discover, characterize, and detect it.
Canonical exemplar target network:
- A 200-employee company with three sites (HQ + two branch offices)
- A public cloud presence (AWS or GCP; simulated in Containerlab + Kubernetes)
- A remote-work VPN population (50 workers)
- WAN underlay using either MPLS or SRv6
- Datacenter at HQ using spine-leaf VXLAN-EVPN
- Internet edge with BGP peering and RPKI deployment
- NSM sensors at each site boundary and cloud egress
Tier 1: Functional Gate
Your project must work before Tier 2 scoring applies.
Gate 1 -- Topology converges:
- Containerlab topology boots without errors
- All routing adjacencies established (show isis/bgp neighbors; no missing peers)
- Loopback reachability across all sites confirmed (ping all loopbacks from HQ)
Gate 2 -- WAN underlay:
- MPLS LSPs or SRv6 SIDs established between all sites
- For SR: at least one SR Policy demonstrated (non-shortest path traffic-engineered)
- traceroute confirms label switching
Gate 3 -- Datacenter fabric:
- Spine-leaf VXLAN-EVPN fabric boots with symmetric IRB
- VM mobility demonstrated: move a container from one leaf to another; reachability maintained
- EVPN Type-2 route update visible after migration
Gate 4 -- RPKI:
- Routinator running and synced
- At least 5 prefixes queried; one Invalid state demonstrated
- FRR configured to reject Invalid routes
Gate 5 -- Protocol RE:
- The unknown protocol's state machine identified with at least 3 named states
- At least 3 message types identified with their field structure
- Protocol identified (or, if novel: documented with a proposed name)
Gate 6 -- NSM detection:
- At least one custom Suricata rule that fires on the unknown protocol's traffic
- The rule tested against a captured pcap of the protocol; confirmed detection
Reports that do not pass all six gates receive no Tier 2 score.
Required Artifacts
| Artifact | Description |
|---|---|
| Containerlab topology YAML | The complete multi-site topology definition |
| FRR/OS configuration files | Per-device configs; should be Ansible/Nornir generated |
| Live topology demonstration | Screencast or live demo during grade session |
| RPKI deployment evidence | Routinator status + FRR rejection of Invalid route |
| Protocol RE write-up | Hex-level dissection, state machine diagram, message type table |
| Custom Suricata rule | .rules file; tested against protocol pcap |
| Capstone report | 35-50 pages; see below |
Capstone Report Structure
Section 1: Network Architecture Design (8-12 pages)
- Design choices and rationale: why MPLS or SRv6 for WAN underlay? Why eBGP unnumbered in the fabric? Why symmetric IRB?
- Site-by-site architecture description
- BGP topology: peering relationships, RPKI deployment
- NSM coverage: sensor placement and coverage rationale
- Threat model: what does this design defend against? (Name at least 3 specific attack classes)
Section 2: Topology Bring-Up Procedures (5-8 pages)
- Step-by-step procedure for cold-starting the topology
- Verification checklist: how to confirm each component is operational
- Troubleshooting guide: the 5 most common failure modes and how to diagnose them
Section 3: Protocol Reverse Engineering (8-12 pages)
- Capture methodology: how you isolated the unknown protocol's traffic
- Hex-level frame dissection: annotated byte captures of at least 5 message types
- State machine diagram with at least 3 states and named transitions
- Protocol identification or proposed specification
- Confidence assessment: what aspects are certain vs. inferred?
Section 4: NSM Coverage and Detection (5-8 pages)
- Suricata rule for the unknown protocol (with rationale for each rule option)
- Detection demonstration: the rule firing against a captured pcap
- Zeek script (optional but valued): behavioral detection of the protocol
- Coverage gap analysis: what aspects of the protocol does the rule NOT detect?
Section 5: Day-2 Operational Runbooks (5-8 pages)
- Adding a new site to the network (step-by-step)
- Responding to a BGP prefix hijack (detection + mitigation)
- Responding to a failed spine switch
- Rolling a new Suricata rule update to all sensors
- Recovering from a VTEP failure in the fabric
Section 6: Limit-of-Defence Statement (2-3 pages) This section must be honest. A limit-of-defence statement identifies what the network design does not protect against.
- Name at least 3 attack classes that this design does not defend against
- Explain the residual risk for each
- Propose what additional controls would be needed to address each
Two-Tier Grading Rubric
Tier 1: Functional Gate
Pass/fail. All six gates must pass. A failing Tier 1 is a non-passing capstone; resubmission is required.
Tier 2: Report Quality (100 pts)
Architecture rationale and integration depth (40 pts)
- Highest score: carrier, datacenter, and adversary-scale decisions are each defended against named alternatives; the capstone reads as one coherent network design rather than three independent pieces
- Middle score: decisions are stated but not defended against alternatives; some integration gaps
- Lowest score: decisions are unexplained; sections feel disconnected
RE methodology and NSM coverage (30 pts)
- Highest score: RE work is systematic (hypothesis → evidence → confirmation cycle); the Suricata rule matches the RE findings; a Zeek behavioral script is provided
- Middle score: RE identifies the protocol but the methodology is not documented; Suricata rule provided without RE-derived rationale
- Lowest score: RE is guesswork; Suricata rule is generic
Operational realism and limit-of-defence honesty (30 pts)
- Highest score: day-2 runbooks match what a network architect actually does at this scale; limit-of-defence statement names specific residual risks and specific mitigations
- Middle score: runbooks are generic checklists; limit-of-defence is vague
- Lowest score: runbooks are theoretical; limit-of-defence says "this design is complete"
B- minimum (70/100 Tier 2) for the NET-301 Certificate of Completion.
Certificate
Passing (all six Tier 1 gates + B- on Tier 2) earns the VCA-NET-301 Certificate of Completion. Combined with NET-101 + NET-201 + NET-301, the student is positioned to sit Cisco CCNP-Enterprise or CCNP-Service-Provider within four months.