Web server benchmarking

Short description: Estimation of web server performance

Web server benchmarking is the process of estimating a web server performance in order to find if the server can serve sufficiently high workload.

Key parameters

The performance is usually measured in terms of:

• Number of requests that can be served per second (depending on the type of request, etc.);
• Latency response time in milliseconds for each new connection or request;
• Throughput in bytes per second (depending on file size, cached or not cached content, available network bandwidth, etc.).

The measurements must be performed under a varying load of clients and requests per client.

Tools for benchmarking

Load testing (stress/performance testing) a web server can be performed using automation/analysis tools such as:

• Apache JMeter, an open-source Java load testing tool
• ApacheBench (or ab), a command line program bundled with Apache HTTP Server
• Siege, an open-source web-server load testing and benchmarking tool
• Wrk, an open-source C load testing tool
• Locust, an open-source Python load testing tool

Web application benchmarks

Web application benchmarks measure the performance of application servers and database servers used to host web applications. TPC-W was a common benchmark emulating an online bookstore with synthetic workload generation.

External links

• Microsoft's patterns & practices Performance Testing Guidance for Web Applications
• Stress tools to test your Web server at Microsoft's knowledge base
• Open Source Performance Testing tools
• SPECweb2009 Release 1.20 Benchmark Design Document
• HTTP(S) benchmark tools, testing/debugging
• h2load - HTTP/2 benchmarking tool - HOW-TO