⚡
Infrastructure disruption
Grid loss, intermittent networking, harsh environments, and limited field maintenance can turn availability into a security property.
M.S. CYBERSECURITY CAPSTONE • 2026–2027
Cyber-Resilient Autonomous Micro Data Centers
A reference architecture for zero-trust operation, physical tamper detection, energy resilience, and verifiable workload governance when traditional data-center assumptions no longer hold.
01 / Research Problem
What happens when the cloud cannot be assumed?
Most security architectures assume dependable power, network connectivity, centralized identity, and physical protection. Autonomous micro data centers challenge each assumption.
🛡️
Expanded attack surface
Remote deployments create combined cyber, physical, supply-chain, identity, and operational risks that cannot be treated independently.
🔎
Evidence gap
Disconnected systems still need to prove what ran, who authorized it, what changed, and whether security controls remained effective.
Primary research question
How can an autonomous micro data center maintain trustworthy computing, cyber resilience, and accountable workload governance during prolonged loss of trusted infrastructure?
02 / Reference Architecture
Resilience by design, not by assumption.
The proposed architecture treats identity, workloads, energy, physical state, and evidence as one continuously evaluated trust system.
TRUST COREPolicy + EvidenceLocal decision authority
IdentityShort-lived credentials
local authentication
local authentication
WorkloadsSigned artifacts
allow-listed execution
allow-listed execution
PhysicalTamper telemetry
environment sensors
environment sensors
EnergyPower state
graceful degradation
graceful degradation
NetworkZero-trust segmentation
disconnected mode
disconnected mode
CloudAWS synchronization
remote governance
remote governance
Zero-Trust Operation
Authenticate and authorize every actor, service, and workload without depending on a permanently reachable control plane.
Tamper Awareness
Use enclosure, motion, environmental, and hardware-state telemetry to detect physical compromise and trigger policy responses.
Energy Resilience
Make compute placement, workload priority, shutdown behavior, and recovery responsive to energy availability.
Verifiable Governance
Create durable evidence of configuration, execution, identity, control state, exceptions, and post-event reconstruction.
03 / Threat Model
Cyber + physical + operational risk.
ThreatFailure ModeDesign Response
Loss of WAN / cloud control planeIdentity, policy, logging, and orchestration become unavailableLocal trust anchor, cached policy, append-only evidence, deferred synchronization
Physical access or enclosure tamperingDevice theft, secrets exposure, sensor or storage compromiseTamper telemetry, encryption, sealed secrets, policy-triggered containment
Energy scarcityUnplanned shutdown, corrupted workloads, loss of monitoringEnergy-aware scheduling, prioritized workloads, controlled degradation
Credential compromiseUnauthorized administrative or workload actionsShort-lived credentials, least privilege, workload identity, offline revocation strategy
Supply-chain compromiseUntrusted firmware, images, libraries, or hardwareSigned artifacts, measured boot, provenance records, component inventory
04 / Experimental Validation
Prove the architecture under failure.
The capstone is organized around repeatable scenarios that create measurable evidence rather than relying only on conceptual analysis.
1
Baseline
Document normal state, identity, configuration, workload provenance, network behavior, and energy profile.
2
Disrupt
Introduce WAN loss, power constraints, physical tamper events, invalid workload artifacts, and compromised credentials.
3
Observe
Measure detection, containment, policy continuity, workload survival, recovery time, and evidence completeness.
4
Verify
Determine whether an independent reviewer can reconstruct what happened and validate that controls behaved as intended.
Local Lab
Mini PCs / SBCs, managed switch, sensors, UPS or DC power, containerized workloads, local identity, and tamper inputs.
AWS Control Plane
Cloud-based policy, logging, storage, inventory, remote attestation workflows, recovery, and synchronization testing.
Field Prototype
Portable enclosure with 3D-printed components, sensor mounting, environmental monitoring, and power-aware operation.
05 / Capstone Artifacts
Research that leaves visible evidence.
01
Reference Architecture
System components, trust boundaries, control layers, data flows, failure states, and operating modes.
02
Threat Model & Risk Register
Cyber, physical, supply-chain, energy, identity, and operational threats tied to mitigations and residual risk.
03
Prototype & Test Harness
Repeatable lab environment for disruption scenarios, telemetry capture, workload governance, and recovery.
04
Evidence Package
Signed logs, configurations, workload provenance, control observations, experiment results, and validation records.
05
Capstone Paper
Problem, literature review, methodology, architecture, experiments, findings, limitations, and future research.
06
Public Research Website
A living record of architecture decisions, experiments, diagrams, demonstrations, publications, and lessons learned.
06 / Intended Contribution
A practical security model for computing beyond dependable infrastructure.
Academic
A defensible graduate research contribution connecting cyber resilience, zero trust, edge systems, and evidence.
Technical
A reproducible architecture and test methodology that others can implement and challenge.
Operational
A blueprint for small, autonomous compute systems that must continue operating under infrastructure loss.
Professional
A portfolio centerpiece demonstrating architecture, cybersecurity, AWS, experimentation, and research capability.
Research Collaboration
research@autonomousmicrodatacenter.com
Follow the build. Challenge the assumptions.
This project is intended to evolve through experiments, peer review, prototype iterations, and published evidence.