Connection Pooling vs Non-Pooling Benchmark

This repository contains two benchmarks to demonstrate the difference in performance between using a connection pool and not using one when connecting to a YugabyteDB instance.

Repository Structure

connection-pooling/
├── connection-non-pool/
│   ├── benchmark_non_pool.go        # Benchmark code without connection pooling
│   ├── Docker-compose.yaml          # Docker Compose file to spin up YugabyteDB
│   ├── go.mod                       # Go module file
│   ├── go.sum                       # Go dependencies file
│   └── vendor/                      # Vendor directory for dependencies
├── connection-with-pool/
│   ├── benchmark_with_pool.go       # Benchmark code with connection pooling
│   ├── Docker-compose.yaml          # Docker Compose file to spin up YugabyteDB
│   ├── go.mod                       # Go module file
│   ├── go.sum                       # Go dependencies file
│   └── vendor/                      # Vendor directory for dependencies
└── README.md                        # This README file

Prerequisites

• Go 1.16 or later installed
• Docker installed and running
• pgcli installed (optional, for connecting to YugabyteDB via CLI)

Getting Started

Step 1: Start YugabyteDB with Docker Compose

cd connection-non-pool # or connection-with-pool
docker-compose up -d

This command will start the YugabyteDB instance in the background.

Step 2: Run the Benchmarks

Without Connection Pooling

Navigate to the connection-non-pool directory and run the benchmark:

cd connection-non-pool
go mod tidy
go run benchmark_non_pool.go

With Connection Pooling

Navigate to the connection-with-pool directory and run the benchmark:

cd connection-with-pool
go mod tidy
go run benchmark_with_pool.go

Step 3: Review the Results

Compare the output of both benchmarks. You should see a significant difference in the number of successful connections and the time taken to complete all requests.

Example Output

Without Connection Pooling:

Benchmark Complete (No Pooling)
Total Requests: 1000
Successful Connections: 170
Connection Failures: 823
Time Taken: 1.0627465s
Average Time per Request: 1.0627465ms

With Connection Pooling:

Benchmark Complete (With Pooling)
Total Requests: 1000
Successful Connections: 999
Connection Failures: 0
Time Taken: 2.142920917s
Average Time per Request: 2.14292ms

Step 4: Tear Down the Environment

To stop and remove the YugabyteDB containers, run:

docker-compose down

Conclusion

This repository demonstrates the performance benefits of using connection pooling when working with a database like YugabyteDB. Without pooling, the application struggles to maintain a high number of concurrent connections, resulting in failures and longer execution times. With pooling, the application can efficiently reuse connections, leading to higher success rates and faster execution.

License

This project is licensed under the MIT License - see the LICENSE file for details.