Rukmini Nazre

Rukmini Nazre

MS in Computer Science student at UMass Amherst

ApexKey: High-Performance Persistent KV Store

Published:

Technologies: C++20, gRPC, Protocol Buffers (Protobuf), CMake, vcpkg, Concurrency (Mutex/RAII)

View on GitHub

Description

  • High-Performance RPC: Engineered a distributed storage engine using C++20 and gRPC, leveraging Protocol Buffers for efficient, low-latency cross-process communication.
  • Concurrency & Memory Safety: Implemented a thread-safe architecture utilizing std::mutex and RAII-style locking, ensuring data integrity across concurrent client requests.
  • Durability Layer: Designed and implemented a Write-Ahead Log (WAL) strategy to persist in-memory state to disk, ensuring 100% data durability across server restarts.
  • Benchmarking & Optimization: Developed a custom benchmarking suite to validate system performance, achieving a throughput of 3,046 ops/sec with an average latency of 0.328 ms on ARM64 architecture.
  • Modern Systems Engineering: Managed complex dependencies via vcpkg and optimized the build lifecycle with CMake, following industry-standard C++ software patterns.