Network administrators often rely on whitepapers, vendor documentation, and product specifications to calculate what kind of load their systems can tolerate and what will happen when they become overloaded. But there’s usually a difference between the spec and the real-world behavior in an implementation. Determining how a network infrastructure responds to both normal traffic loads and extreme stress loads is critical.
Unfortunately there are complications that make a simple utilization or bandwidth threshold calculation unreliable. Most notably, many different types of network traffic are commonplace in today’s corporate network. Each traffic type impacts the network in a different way. And since each network is used in a unique way with unique traffic patterns, each must be assessed based on its own type of utilization.
The best way to determine network capacity and fail points is to actually generate traffic in patterns similar to those found in the production network, and then amplify them until something breaks. The two most common ways to do this, network playback and network traffic generation are the subject of this article. You’ll see the benefits and drawbacks of each and how they are used with other tools and tests to validate configurations and discover weak spots in an infrastructure.
You can watch my video: Troubleshooting a Network with Traffic Simulation for more information.
How do you test network capacity?