RTS Questionnaire · Question 48 of 52 · file: answers/q48.html
Have you worked with database scaling strategies such as read replicas, horizontal partitioning/sharding, and caching?
DatabasesScalingCaching
✅ Yes — I apply these scaling strategies together, choosing based on measured bottlenecks rather than preference.
🧪 Interactive proof: this answer has its own page with a live scaling-ladder workflow, the read routing, the cache inventory and the load test that validates the topology behind it — open the Q48 database-scaling proof →
Technical detail
- 📖 Read replicas — routing read-heavy traffic to replicas, tolerating replication lag in appropriate paths (analytics, listings) and pinning critical read-after-write operations to the primary; managed equivalents on RDS/Aurora, Azure SQL and PostgreSQL streaming replication.
- 🧱 Horizontal partitioning / sharding — splitting large tables by time, tenant or hash key so the working set stays manageable; in NoSQL, achieving distribution through partition-key design (DynamoDB, Cassandra/Keyspaces) with hot-partition avoidance and adaptive capacity.
- 🧠 Caching — multi-layer strategy: in-process memoisation, distributed Redis cache with cache-aside and write-through, HTTP/response caching and CDN at the edge; explicit TTL and invalidation contracts, stampede protection with jittered TTLs and single-flight locks, and cache-hit-ratio monitoring.
- 📚 Complementary techniques — CQRS read models and materialised views for expensive aggregations, archival/partition switching for cold data, connection pooling and batching, and search/analytics offloading (Athena, Elasticsearch-style clusters) so the transactional database stays lean.
- 📊 Method — measure first (p95/p99 latency, cache hit ratio, replication lag, lock waits), then scale the actual constraint. Scaling decisions and their trade-offs are documented, and load tests validate the new topology before release.
🧾 CV evidence: CV: AWS database services (RDS, DynamoDB, Keyspaces, DocumentDB) with read-replica and sharding patterns; Redis among core data technologies in the 2025/2026 CV; performance tuning and data-warehouse/ETL experience (Athena, Glue, Redshift).