A fast, lightweight, in-memory key-value store with a straightforward HTTP API, developed using Rust for high efficiency.

• Rust version 1.70 or above
• Cargo package manager

# Download the repository
git clone https://github.com/aannaannyaaa/KV-Cache-Project
cd KV-Cache-Project

# Compile the project
cargo build --release

# Run the server
./target/release/key_value_cache

# Install directly using cargo
cargo install --git https://github.com/Rushhaabhhh/Key-Value-Cache.git

# Run the server
key_value_cache

# Build the Docker image
docker build -t key-value-cache .

# Run the Docker container
docker run -p 7171:7171 key-value-cache

Default parameters can be adjusted by adding a .env file in the root directory:

PORT=7171
MAX_KEY_SIZE=256
MAX_VALUE_SIZE=256

POST /put

Request body example:

{
  "key": "user:1234",
  "value": "John Doe"
}

Success response:

{
  "status": "OK",
  "message": "Key successfully inserted or updated."
}

GET /get?key=user:1234

If found, response example:

{
  "status": "OK",
  "key": "user:1234",
  "value": "John Doe",
  "message": "Key retrieved successfully."
}

If not found:

{
  "status": "ERROR",
  "message": "Key not found."
}

• Employs an LRU cache using Rust's LinkedHashMap to track usage and evict the least recently used entries.
• Monitors system memory continuously, triggering eviction when usage hits 70% capacity.
• Uses a progressive eviction approach that increases eviction batch sizes gradually, preserving responsiveness.

• Built on Tokio's async runtime to handle multiple simultaneous requests efficiently.
• Thread-safe through Rust’s Arc<RwLock<>> for concurrent cache operations.
• Limits key and value sizes to maintain stable performance and prevent abuse.

• Supports graceful termination to avoid data loss on shutdown.
• Provides consistent and clear error responses through Rust’s error handling.
• Monitors resources actively to adapt to workload changes dynamically.

• Designed as a stateless application (outside of cache data) to easily scale horizontally behind a load balancer.
• Optimized for minimal memory usage, suited for deployment in environments with limited resources.

This service delivers exceptional speed:

• Supports over 50,000 requests per second on typical hardware.
• 95th percentile response times below 5 milliseconds.
• Zero failing requests under heavy concurrent load (100+ users).

• Throughput: 51,455 requests/second
• Latency: 95% of requests completed in under 3.14ms
• Reliability: No failures in nearly 17 million requests
• Cache Efficiency: Achieved 100% effective cache hits with over 8 million hits
• Concurrency: Maintained stability with 100 simultaneous virtual users

Would you like the text adjusted to be more formal or casual?