Scaling TKE's Sales Configurator Application
TKElevator @ Infosys
2023 — 2026
The Challenge: During peak usage windows, the enterprise configuration module suffered under massive data processing overhead, severe thread contention, and frequent database timeout spikes that degraded system responsiveness.
⚙️ HLD Optimization
- Offloaded heavy computational workloads to background thread pools, maintaining main thread responsiveness.
- Replaced unindexed queries with compiled stored procedures and targeted covering indexes.
- Integrated distributed Redis caching layers to bypass redundant calculations for frequent request payloads.
⚖️ Architectural Rationale
- Prioritized uninterrupted user experience by decoupling blocking operations from request-response lifecycles.
- Enforced database-level tuning to preserve strict transactional ACID compliance without cache race conditions.
- Leveraged Redis output caching specifically for repetitive configuration patterns, drastically cutting thread wait times.
📊 Quantified Impact
- Validation: Verified via App Insights and SQL Execution Plan telemetry.
- Result: Average API latency dropped from ~340ms to ~190ms (40% faster); system throughput scaled by +50%.