Speed me up!

Speed me Up!Performance Testing Best Practices

Goal: performance tests determine system and user responses, throughput, reliability, scalability of a system under load

“Performance test is easy. It is easy to design not realistic scenarios. It is easy to collect irrelevant data. Even, with a good scenario and good data it is easy to use incorrect statistical Method.” – Alberto Savoia

SCENARIOS: Identify the scenarios that are most commonly executed or most resource-intensive

WORKLOAD MODEL: User Session Duration in average. It is important to define the load levels that will translate into concurrent usage, overslapping users, or user sessions per second.

USER SCENARIO: Navigational Path, including intermediate steps or activities, taken by the user to complete a task. We will call it User Session from now on.

THINK TIMES: User thinktimes Pause between pages during a User Session depending on the User Type*

User Types: Identify the User … new, revisiting… devices, networks…

Performance Acceptance Criteria: Response time, System load, Throughput ...

METRICS: Only well-selected metrics that are analyzed correctly and contextually provide information of value.

DESIGN TEST: Using your scenarios, key metrics, and workload analysis …

RUN TEST: the load simulation must reflect the test design

ANALYZE RESULTS: Find bottlenecks, memory leaks, cpu hogs, bad sofware design …

Monitor Performance in production: and check if the assumptions from your test phase are correct. Adapt your tests always to production experience.

• http://www.raosoft.com/samplesize.html

• http://analyze.websiteoptimization.com/wso

• http://mobitest.akamai.com/m/index.cgi

• http://stevesouders.com/mobileperf/mobileperfbkm.php