Statistics Project

This project provides two thread-safe implementations of the Statistics interface for collecting and calculating statistical data such as minimum, maximum, mean, and variance. The project also demonstrates SPI (Service Provider Interface) to allow switching between different implementations dynamically.

---

Features

• Thread-Safe Implementations:
  • AtomicThreadSafeStatistics: Uses Atomic variables for thread safety.
  • LockBasedThreadSafeStatistics: Uses ReentrantLock for thread safety.
• SPI Integration: Allows dynamic switching between implementations.

---

Project Structure

statistics
├── statistics-spi        # SPI interface module
├── statistics-atomic     # Atomic implementation
├── statistics-lock       # Lock-based implementation
├── statistics-test-base  # Base test cases
├── statistics-client     # SPI client to use implementations
└── dist                           # Packaged distribution

---

Getting Started

Prerequisites

• Java: Ensure JAVA_HOME is set to a valid JDK installation.
• Maven: Verify Maven is installed and available in PATH.

Building the Project

1. Build all modules:

   ./build-and-package.sh

This script builds the project, and required artifacts are placed in dist.

---

Running the SPI Client

1. Execute the client:

   cd ./dist/bin/
   source run-spi-client.sh

---

Development

Adding a New Implementation

1. Create a new module (e.g., statistics-newimpl).
2. Implement the Statistics interface.
3. Register the implementation in META-INF/services/in.shashwattiwari.statistics.Statistics.

Running Tests

To run all tests:

mvn test

---

Project Dependencies

• JUnit 5: Used for testing.
• Maven: For build and dependency management.

---

Future Enhancements

• Add more statistical methods.
• Improve SPI client logging.
• Support for distributed statistics computation.
• Supports performance benchmarking using JMH.
• Can be extended to handle backpressure scenarios.
• Utilise Actor Model based Frameworks like Akka, which inherently handle concurrency.
  • One message at a time get processed by respective actors.
  • Would be easier to scale horizontally.

---

Contact

For questions or feedback, please contact:

• Shashwat Tiwari: [[email protected]]