While preparing for the AWS SAP-C02, many candidates confuse the roles of Application Discovery Service, Systems Manager, Migration Hub, and Trusted Advisor. In the real world, this is fundamentally a decision about data granularity vs. operational overhead during large-scale migration planning. Let’s drill into a simulated scenario.
The Scenario #
GlobalForge Manufacturing operates 47 data centers across APAC and EMEA, hosting approximately 2,200 workloads on a mix of bare-metal servers and VMware ESXi clusters running both Linux (RHEL 7/8, Ubuntu) and Windows Server (2012 R2–2022). The CTO has secured board approval for a 24-month cloud-first migration to AWS.
The migration team must accomplish three critical objectives before selecting migration tooling:
- Deep Workload Profiling: Capture OS configuration, running processes, network dependencies (which apps talk to which databases), CPU/memory utilization patterns over 14 days, and filesystem I/O metrics.
- Application Grouping: Logically cluster interdependent servers (e.g., web tier + app tier + database tier) into migration waves.
- Cost-Optimized EC2 Sizing: Generate instance type recommendations that minimize AWS spend while meeting performance SLAs.
Key Requirements #
Select THREE actions that collectively satisfy all objectives with the lowest operational complexity and highest cost visibility.
The Options #
- A) Deploy AWS Application Discovery Agent on all physical and virtual machines to evaluate existing applications.
- B) Deploy AWS Systems Manager Agent (SSM Agent) on all physical and virtual machines to evaluate existing applications.
- C) Use AWS Systems Manager Application Manager to group servers into logical applications for migration planning.
- D) Use AWS Migration Hub to group servers into logical applications for migration planning.
- E) Generate EC2 instance type recommendations and cost estimates via AWS Migration Hub.
- F) Export server sizing data to AWS Trusted Advisor and apply its cost optimization recommendations.
Correct Answer #
A, D, E
Quick Insight: The FinOps Imperative #
This scenario tests your ability to distinguish purpose-built migration services from general operational tools. The wrong choices (B, C, F) would force you to build custom integrations, manually correlate data, and lack native TCO calculators—adding 40–60 engineering hours per 100 servers. The correct trio (A+D+E) delivers automated dependency mapping, wave planning, and quantified cost projections in a single workflow, critical for securing CFO approval.
💎 The Architect’s Deep Dive: Why Options Fail #
Correct Answer #
Options A, D, and E
Step-by-Step Winning Logic #
This is a professional-level migration discovery problem requiring three distinct capabilities:
-
Deep System Telemetry (Option A)
The AWS Application Discovery Agent is the only tool listed that captures:- Process-level metadata (what services are running)
- Network connection topology (source IP → destination IP/port pairs)
- Time-series performance metrics (CPU, RAM, disk I/O, network throughput)
- Filesystem and configuration data
This data is essential for dependency mapping and right-sizing. SSM Agent (Option B) is designed for operational management (patching, command execution), not discovery.
-
Application Grouping (Option D)
AWS Migration Hub acts as the central orchestration plane, allowing you to:- Visualize discovered servers
- Manually or automatically group them into applications (e.g., “SAP ERP Production”)
- Track migration status across waves
Systems Manager Application Manager (Option C) is for post-migration operational monitoring, not pre-migration planning.
-
Cost-Optimized Sizing (Option E)
Migration Hub integrates with AWS Migration Evaluator and uses Application Discovery data to:- Recommend EC2 instance families (e.g., m6i.xlarge vs. c6i.2xlarge)
- Project monthly costs for each application group
- Compare on-premises TCO vs. AWS TCO
Trusted Advisor (Option F) provides post-deployment cost optimization (e.g., “delete unused EBS volumes”), not pre-migration sizing.
The combination A+D+E creates a closed-loop discovery → grouping → costing workflow. #
The Traps (Distractor Analysis) #
-
Why not B (SSM Agent)?
SSM Agent requires EC2 instances (or hybrid activations) and doesn’t collect network topology or process-level data. It’s the wrong tool for discovery—you’d use it after migration for patch management. -
Why not C (Systems Manager Application Manager)?
Application Manager is for runtime observability (CloudWatch dashboards, OpsCenter integration). It can’t ingest Application Discovery data or group on-premises servers. -
Why not F (Trusted Advisor)?
Trusted Advisor operates on existing AWS resources. It can’t analyze on-premises servers, and manually exporting data to it would require custom scripting with no TCO comparison features.
The Architect Blueprint #
graph TD
A[On-Premises Servers
2,200 Linux/Windows] -->|Deploy| B[Application Discovery Agent]
B -->|Stream Metrics| C[Application Discovery Service]
C -->|Export Data| D[AWS Migration Hub]
D -->|Group Servers| E[Application Groups
Wave 1: SAP, Wave 2: CRM]
D -->|Analyze Utilization| F[Migration Hub Recommendations]
F -->|Output| G[EC2 Instance Types + Monthly Cost]
style B fill:#FF9900,stroke:#232F3E,color:#FFF
style D fill:#FF9900,stroke:#232F3E,color:#FFF
style G fill:#3F8624,stroke:#232F3E,color:#FFF
Diagram Note: The Application Discovery Agent feeds granular telemetry into Migration Hub, which clusters servers into application groups and generates cost-optimized EC2 sizing recommendations—creating a traceable decision path from discovery to budget approval.
The Decision Matrix #
| Option | Est. Complexity | Est. Cost Impact | Pros | Cons |
|---|---|---|---|---|
| A: Application Discovery Agent | Medium (agent deployment at scale) | Low ($0.00 for agent; $0.0011/GB data ingested) | • Process/network topology • 14-day performance baselines • Native Migration Hub integration |
• Requires installation on every host • Firewall rules for outbound HTTPS (443) |
| B: SSM Agent | Medium | N/A (wrong tool) | • Useful post-migration for patching | • No dependency mapping • No TCO calculator integration |
| C: Systems Manager App Manager | Low (UI-based) | N/A (wrong tool) | • Great for post-migration ops | • Can’t ingest on-prem discovery data |
| D: Migration Hub | Low (managed service) | $0 (console is free) | • Central migration dashboard • Application grouping UI • Integration with cost tools |
• Requires data from Discovery Agent |
| E: Migration Hub Recommendations | Low (auto-generated) | Potential 15–30% savings vs. lift-and-shift | • Quantified EC2 cost projections • Family/size optimization |
• Recommendations assume steady-state utilization |
| F: Trusted Advisor | High (manual export) | N/A (wrong tool) | • Useful for post-deployment checks | • No on-prem data ingestion • No migration-specific features |
FinOps Insight: For a 2,200-server estate, Application Discovery Agent data ingestion costs ~$500–$800 over 14 days, but the resulting right-sizing recommendations typically yield $50K–$150K/year in EC2 savings (based on AWS case studies showing 20–40% reduction vs. 1:1 lift-and-shift).
Real-World Practitioner Insight #
Exam Rule #
For SAP-C02, if you see “capture network connections and processes” + “group for migration” + “cost-optimized EC2 sizing,” always select:
- Application Discovery Agent (data collection)
- Migration Hub (orchestration)
- Migration Hub cost features (TCO analysis)
Real World #
In production, you’d also:
- Use Agentless Discovery Connector for VMware environments where agent deployment is restricted (trades granularity for speed)
- Integrate with AWS Application Migration Service (MGN) for the actual replication/cutover
- Validate recommendations with Well-Architected reviews—Migration Hub might suggest m6i.xlarge based on average CPU, but bursty workloads might need c6i instances
- Run proof-of-concept migrations for Tier-1 apps before committing to instance families
The exam simplifies the workflow; real migrations require pilot waves, rollback plans, and post-migration optimization sprints.