Queueing and Connection Limits
You'll learn how browsers restrict the number of simultaneous connections per domain and how this creates queueing delays that slow down page loads.
The Browser's Traffic Control System
Browsers don't download everything at once—they impose connection limits per domain to prevent overwhelming servers and manage network resources efficiently. In HTTP/1.1, most browsers allow only 6 concurrent connections per domain. In HTTP/2, this shifts to a single multiplexed connection, but the principle of managed concurrency remains.
Think of it like checkout lanes at a grocery store. If you have 20 customers (requests) but only 6 lanes (connections), the remaining 14 customers must queue and wait their turn. This waiting time is queueing delay.
How Queueing Impacts Performance
When you examine a waterfall chart, you'll notice some requests show a gray "queueing" bar before they even start downloading. This happens when:
- More resources exist than available connections
- Earlier requests haven't completed yet
- The browser is prioritizing other critical resources
For example, if your page loads 30 images from the same domain over HTTP/1.1, only 6 download simultaneously. The other 24 sit in a queue, adding precious milliseconds (or seconds) to your total load time.
Connection Overhead Compounds the Problem
Each queued request must wait for an available connection and then go through connection establishment (which you learned about in "Connection Overhead and Timing"). This creates a multiplicative delay effect—the later a resource appears in the queue, the longer users wait.
Key Takeaway: Browsers limit concurrent connections per domain (typically 6 for HTTP/1.1), causing later requests to queue and wait, which significantly increases total page load time—understanding this helps you optimize resource loading strategies like domain sharding or HTTP/2 adoption.