FTCache: Fault-Tolerant Distributed Cache for HPC Systems

Overview

FTCache is a fault-tolerant, distributed caching layer designed for high-performance computing (HPC) environments. It leverages consistent hashing and a hash ring architecture to provide load balancing and resilience against node failures. The system is implemented in C++ and supports integration with local and parallel file systems.

Features

• Consistent Hash Ring for dynamic node management
• Fault tolerance with automatic failover
• Load balancing across multiple cache nodes
• Support for both local and parallel file systems (PFS)
• Performance benchmarking and result visualization scripts

Project Structure

.
├── src/                # Source code (C++ files, hash ring, main entry points)
│   ├── ConsistentHashRing.hpp
│   ├── v1_main.cpp ... v6_main.cpp
│   ├── Local_Cache/    # Simulated local cache nodes
│   └── PFS_Source/     # Parallel file system source files
├── Archive/            # Old presentations, scripts, and results
├── build/              # Build artifacts and CMake files
├── Presentations/      # Project presentations and scripts
├── Paper/              # Research paper drafts and references
├── README.md           # This file
├── CMakeLists.txt      # CMake build configuration
└── ...                 # Other supporting files

Getting Started

Prerequisites

• C++17 compatible compiler (e.g., g++, clang++)
• CMake 3.10+
• Python 3.x (for plotting and analysis scripts)
• (Optional) Docker (for containerized node simulation)

Build Instructions

1. Clone the repository:

   git clone https://github.com/seunshix/replicated-cache-layer.git
   cd FTCache

2. Build with CMake:

   mkdir -p build
   cd build
   cmake ..
   make

3. Or, build a specific version directly:

   g++ -std=c++17 -o v1_main ../src/v1_main.cpp

Running Tests

• To run the main cache simulation:

  ./build/FTCache

• To run a specific version:

  ./src/v2_main

Plotting Results

• Use the provided Python scripts to visualize results:

  python3 src/plotting_results.py

Usage

• Configure cache nodes and PFS sources in the src/Local_Cache/ and src/PFS_Source/ directories.
• Modify or extend the hash ring logic in src/ConsistentHashRing.hpp.

Documentation

• See the Presentations/ and Paper/ folders for methodology, results, and research context.
• Refer to code comments in src/ for implementation details.

Contributing

Contributions are welcome! Please open issues or submit pull requests for improvements or bug fixes.

License

This project is for academic and research purposes.