AWS Database study plan
AWS Database · Give each week a clear focus.
8 weeks · About 6 hours per week
Workload-Specific Database Design
Classify a set of 12 sample workloads by access pattern (point lookup, range scan, aggregate, graph traversal, time-series append) before choosing any engine
Distinguish OLTP vs OLAP vs streaming-ingest requirements and translate each into concrete sizing and concurrency constraints
Decompose each workload into read/write ratio, item size, latency target, and consistency requirement
Identify where normalization vs denormalization is forced by the engine's data model
Your checkpoint
A written classification table covering 12 workloads with access pattern, latency target, consistency model, and first-choice engine family (null + flagged where uncertain).
Workload-Specific Database Design
Compare relational (RDS/Aurora), key-value (DynamoDB), document (DocumentDB), wide-column (Keyspaces), graph (Neptune), in-memory (ElastiCache/MemoryDB), time-series (Timestream), and warehouse (Redshift) engines on partitioning, scaling model, and query flexibility
Design a DynamoDB partition key and secondary index set for two access patterns, then predict hot-partition risk
Model the same schema relationally with indexes and compare query cost against the NoSQL alternative
Select read replicas vs caching vs sharding for a stated read-heavy scenario and justify the tradeoff
Your checkpoint
An engine decision matrix scoring 8 engines against 5 criteria (latency, scale model, query flexibility, operational effort, cost shape) plus 2 worked schema designs.
Workload-Specific Database Design & Deployment and Migration
Resolve 10 design tradeoff scenarios (consistency vs latency, cost vs durability, managed vs self-managed) and defend each choice
Map each workload from week 1 to a deployment topology (single-AZ, Multi-AZ, multi-region, global table)
Contrast homogeneous vs heterogeneous migration paths and pick the required tooling for each
Estimate downtime and cutover risk for a chosen migration approach
Your checkpoint
A design-to-deployment mapping document for all 12 workloads, each with a named migration path and an estimated downtime window.
Deployment and Migration
Sequence a full migration: assessment, schema conversion, data transfer, validation, cutover, rollback
Configure a migration approach using native backup/restore, snapshots, or logical replication and state when each is preferred
Design a data validation strategy that compares row counts, checksums, and sampled records
Write the rollback and cut-forward criteria for a production cutover
Your checkpoint
A 1-page cutover runbook with ordered steps, validation queries, an explicit rollback trigger, and a go/no-go checkpoint.
Management and Operations
Provision a database through infrastructure-as-code and separate configuration from secrets
Plan maintenance windows, patching cadence, and parameter-group change propagation for a live workload
Design a backup and restore policy that meets a stated RPO/RTO, then test an actual restore
Automate a routine operational task (snapshot cleanup, index rebuild, scaling action) and verify it runs unattended
Your checkpoint
A restore test record showing measured restore duration compared against a stated RTO, plus the automation script and its last successful run.
Monitoring and Maintenance
Instrument a database with the right metrics for CPU, memory, connections, storage, replication lag, and throttling
Diagnose three failure signatures (connection exhaustion, I/O saturation, hot partition or write throttling) from metric patterns alone
Configure slow-query and audit logging, then extract the top 5 costliest statements from the output
Set alert thresholds that distinguish actionable signals from noise
Your checkpoint
A monitoring runbook mapping 5 symptom patterns to the metric graph, root cause, and first remediation step, with alert thresholds defined.
Security and Compliance
Design least-privilege access using IAM roles, database-native users, and network controls layered together
Configure encryption at rest and in transit with correct key rotation and grant boundaries
Place credentials in a managed secret store and prove the application never embeds them
Map a workload to a compliance regime and list the audit-logging and retention controls it demands
Your checkpoint
A security control checklist for one workload covering identity, network, encryption, secrets, and audit logging, with each item marked implemented or gap.
Workload-Specific Database Design & Deployment and Migration & Management and Operations & Monitoring and Maintenance & Security and Compliance
Re-rank your weakest domains from prior milestones and rebuild the two thinnest areas
Work mixed-domain scenarios that combine design, migration, and monitoring in a single question
Time-box a full-length practice set and categorize every miss by domain and by failure type (knowledge gap vs misread)
Convert each recurring miss into a single corrective note you can review in the last 48 hours
Your checkpoint
A scored practice set with all misses tagged by domain and cause, plus a final one-page review sheet of no more than 20 corrective notes.
Milestones
Week 2: Can justify an engine choice for 8 of 8 engine families from the decision matrix in under 60 seconds each, with no more than 1 flagged as uncertain.
Week 4: Cutover runbook executes end-to-end in a test environment with validation queries passing and a successful simulated rollback.
Week 6: Correctly identify root cause for 3 of 3 injected failure signatures using metrics alone, before consulting logs.
Week 8: Practice-set score at or above target in every domain, with zero domains below the passing threshold and all misses reduced to a 20-note review sheet.