Connection Pooling

A technique where a set of pre-established database connections are reused rather than creating a new connection every time a client makes a request.

The Problem It Solves

Creating a new database connection is expensive — it involves network handshake, authentication, memory allocation, and session setup. If every request creates and destroys a connection, it leads to high latency and resource wastage under heavy load.

How It Works

Clients            Connection Pool          Database
  |                      |                      |
  |---Request----------->|                      |
  |                      |---Use Connection---->|
  |                      |<--Result--------------|
  |<--Response-----------|                      |
  |     (connection returned to pool, not closed)
  |---Request----------->|                      |
  |     (reuses same connection)

Key Parameters

Parameter	Description
Minimum Pool Size	Connections kept open even when idle
Maximum Pool Size	Max connections allowed in the pool
Connection Timeout	How long a client waits for a free connection
Idle Timeout	How long an unused connection stays in the pool
Max Lifetime	Maximum age of a connection before it's recycled

Pool Size Rule of Thumb (HikariCP)

Pool Size = (Core Count × 2) + Effective Disk Spindles

Important: Bigger pool ≠ Better performance. Too many connections can overwhelm the database.

Connection Pool States

• Idle → Available, waiting to be borrowed
• Active → Currently in use by a client
• Closed → Removed from pool (timed out or errored)

Benefits vs Drawbacks

Benefits	Drawbacks
Reduced connection overhead	Consumes memory even when idle
Faster response times	Misconfigured pool size can cause issues
Better resource management	Stale connections can cause errors
Handles traffic spikes better	Adds complexity to the system

Popular Connection Pool Libraries

Language	Library
Java	HikariCP, c3p0, DBCP
Python	SQLAlchemy Pool, psycopg2 pool
Node.js	pg-pool, mysql2 pool
.NET	Built-in ADO.NET pooling