FluxoBid: High-Performance OpenRTB 2.5 Core

A high-performance, asynchronous Real-Time Bidding (RTB) engine built with C++20. Designed for ultra-low latency (<1ms internal) and high-concurrency environments, this project implements a bidding core capable of processing hundreds of OpenRTB 2.5 requests per second.

Key Architectural Features

• Zero-Copy JSON Ingestion: Leverages std::string_view and yyjson to map OpenRTB requests directly to C++ structs without heap allocations or redundant data copies.
• Hybrid C++20 Module Strategy: Utilizes a robust architectural split with Named Modules for business logic and Global Module Fragments to maintain stability with third-party networking headers (Asio).
• Early-Exit Validation: A "Pre-Flight" layer identifies "deal-breakers" (missing IDs, low timeouts) to trigger instant HTTP 204 No Content responses, preserving CPU for high-value auctions.
• Type-Safe Modeling: Employs C++20 Concepts and std::optional to enforce strict schema adherence for complex nested objects like Device, Site, and Banner configurations.
• Asynchronous I/O Stack: Implements a high-concurrency Server/Session pattern using Asio Coroutines to manage thousands of simultaneous TCP connections.

Tech Stack & Requirements

• Language: C++20 (utilizing Concepts, and Coroutines)
• Compiler: Clang 16+ (optimized for diagnostics and C++20 standard compliance)
• Build System: CMake 3.28+
• JSON Engine: yyjson (C-based for maximum performance)
• Networking: Asio (Standalone version)
• Testing: Catch2 v3

Project Structure

FluxoBid-Processor/
├── include/                # Public API, Models, and Concepts
├── src/
│   └── core/               # Bidder, Engine, and CampaignStore logic
│   └── network/            # Server and Session connection management
├── tests/                  # Unit Test suite (Catch2) verifying OpenRTB compliance
│   ├── test_bidder.cpp     
│   └── test_engine.cpp     
├── CMakeLists.txt          # Modern CMake build with C++20 Module support
└── main.cpp                # Application entry point & Asio event loop

Current Progress: Verified Parsing & Unit Testing

The project has successfully reached the Ingestion & Decisioning Milestone, moving from raw network bytes to a validated bidding decision.

• OpenRTB Ingestion & Auction Logic: Implemented a "Sieve" matcher that parses multi-impression OpenRTB 2.5 payloads and filters campaigns based on Geography, Size, and Price Floor.
• Asynchronous Network Core: Developed a custom Server/Session architecture using Clang 16+ and Asio to handle concurrent TCP connections with minimal overhead.
• Protocol Handling: Supports the standard OpenRTB 2.5 response lifecycle, including automated HTTP 204 "No-Bid" signals and dynamic JSON BidResponse generation via yyjson.
• Automated Verification: Core logic is verified via a Catch2 v3 test suite (test_bidder) that validates stable field mapping, campaign matchmaking, and memory safety across the pipeline.