One thing to keep in mind is it is easy to be led astray about your scalability with Apache Bench. As one example, a repeated get of the same resource is often effortless for an application server due to caching, whereas a real load will be asking for many different uncached resources.

Also, the article does a good job covering concurrency. That -c argument is extremely important, and if you haven't set it correctly, your test is probably not giving you a "real" result.

ab is a simple load testing command from Apache. It benchmarks your HTTP server by automating a scenario of sending multiple requests with concurrent clients.