Design and Implementation of a Mini Database Connection Pool

Introduction

The basic idea of a database connection pool is to maintain a buffer of pre‑created database connections that can be borrowed and returned, reducing the overhead of frequent connection creation and release, which is crucial for high‑QPS distributed systems.

Key Design Thoughts

1. The pool stores and manages database operation pipelines, not just holds them. 2. It should be initialized via external configuration files. 3. Concurrency must be considered to allocate pipelines safely under multithreaded access. 4. Pipelines should be reusable; returning them marks them as available rather than truly closing them.

Mini Pool Implementation Overview

The implementation defines several core interfaces and classes:

IMyPool : Provides basic pool services such as obtaining a connection.

MyDefaultPool : Implements IMyPool and holds a collection of connections, using a thread‑safe Vector.

MyPooledConnection : Wraps a JDBC Connection, tracks an isBusy flag, and overrides close to merely mark the connection as free.

MyPoolFactory : A singleton factory that supplies the IMyPool implementation.

DBConfigXML : Represents external XML configuration for driver loading and pool parameters.

Initialization

During startup, the pool reads the XML configuration, loads the JDBC driver, and creates the initial set of connections according to the configured pool size.

Creating a Pooled Connection

The pool checks whether the maximum number of connections has been reached; if not, it creates a new MyPooledConnection and sets its isBusy flag accordingly.

Obtaining a Pooled Connection

If no free connection is available, the pool attempts to create a new one; otherwise it returns an existing idle connection, marking it as busy.

Retrieving the Real JDBC Connection

The method is synchronized to avoid race conditions, validates that the underlying JDBC connection has not timed out, and ensures the isBusy flag is correctly set.

Factory and Testing

MyPoolFactory provides a global access point to the pool instance, and a simple Test class demonstrates acquiring and releasing connections, showing the expected output.

Conclusion

The article walks through the complete design and a functional mini‑implementation of a database connection pool, illustrating essential concepts such as resource reuse, thread safety, configuration‑driven initialization, and practical testing.