Comparison of HTTP vs. Browser Load Tests

Criteria HTTP Real Browser
User simulation No client side rendering Real user simulation
Script implementation and customization Difficult when web sites are complex Simple scripts, easy to customize using EveryStep recorder
Script replay Low level analysis required You see what you get using EveryStep
Script maintainability Programming skills required Easy, as almost all done by EveryStep recorder
Multi Browser Support Can emulate web browser using Agent-string Chrome and Internet Explorer
Footprint on load injection machine Low, up to 500-100 sessions per load injector High, up to 8-12 sessions per load injector
Recommended for DevOps Depends on actual test scenario Yes, easy to use and realistic figures
Recommended for Load Tests • Webservers
• Web APIs: JSON, SOAP, Rest, XML.
• Single web pages without JS or CSS
• Web Applications.
• Compex pages containging Java Scripts and CSS.
Pro’s and Con‘s Pro‘s
Low Cost
Exact HTTP/S targetCon‘s
No scripting
No JS or CSS
Pro’s
Real User Simulation
Get what you see in browser exactly
Support for JS, CSS, RIACon’s
High Cost
Lot‘s of data
Costs Low High