AWS SDE Internship
EC2 Network Interface Query System
Designed and built a GraphQL-based proof of concept for querying EC2 network interface data, reducing payload size by 80% and achieving sub-second latency.
August 1, 2025
backendgraphqlawsfull-stack
01
Context
During my SDE internship at AWS on the EC2 ZIM Core Team in Arlington, VA, I worked on improving how teams query EC2 network interface data. The existing REST-based approach was inefficient, returning large payloads when clients often needed only a subset of fields.
02
What I Built
A GraphQL-based proof of concept that enables selective field retrieval for EC2 network interface data. The service includes a dynamic SQL query builder with optimized joins, caching strategies, and was containerized with Docker via RDE. Collaborated with the cZIM team for validation and integration.
03
Key Decisions
1Chose GraphQL over REST for flexible, client-driven queries
2Built dynamic SQL query builder eliminating N+1 query issues
3Implemented caching strategies to improve database throughput by 25%
4Containerized the service with Docker via RDE for easy deployment
5Designed within AWS Coral framework for distributed microservice compatibility
04
Challenges
→Ensuring system reliability and maintainability at AWS scale
→Benchmarking resolver performance under high concurrency
→Integrating with existing AWS infrastructure and teams
05
Outcomes
✓80% reduction in payload size for complex queries
✓Sub-second latency achieved for production-scale queries
✓Sustained 330+ requests/sec with zero data loss in benchmarks
✓Design adopted for large-scale production integration
06
Tech Stack
GraphQLDockerAWS CoralSQLJava