Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

README.md

cpp-compute Service

This is the C++ Compute Node for the Library System, responsible for heavy computational tasks such as image hashing, perceptual hash extraction, and anomaly detection in returned items.

Architecture & Dependencies

  • Framework: gRPC (C++)
  • Concurrency: Thread-per-request (or async depending on gRPC setup)
  • Hardware Dependency (Current): None. The current implementation uses CPU-bound cryptographic hashing (SHA-256 via OpenSSL) as a placeholder for the perceptual hash baseline.
  • Hardware Dependency (Target Phase 7): Once anomaly detection and true perceptual tensor hashing are introduced, this node will require GPU acceleration (NVIDIA CUDA, Compute Capability 7.0+). Running the future version on a CPU-only host will cause performance degradation.

Future Plans (Phase 7)

In Phase 7, this service will be integrated directly into the deployment pipeline with automated test coverage enforcement. Code coverage metrics (C++) will be collected alongside C# metrics. Furthermore, the hashing logic will be replaced with a real ML tensor model for anomaly detection on book images, at which point the CUDA hardware requirement will take effect.

Security Notes

  • Inputs over 5MB are rejected to prevent resource exhaustion.
  • Image IDs are sanitized before logging to mitigate CRLF log injection vulnerabilities.