CT-PT: Certified Tester Performance Testing

Question

You are overseeing testing for an existing distributed system that manages inventories of automobile and light-truck tires from multiple warehouses throughout the country. The system is being enhanced to track incoming restocking shipments when they enter the warehouse and outbound sales shipments when they leave the warehouse, all in real time. System load has traditionally peaked on Mondays because of demand accumulated over the preceding weekend.

The current production platform has been fully upgraded, but it is experiencing reporting-performance problems that were not observed during testing. Your analysis shows that the test system has equivalent hardware, a substantially smaller database, and was tested with the same number of virtual users as the number of actual production users.

Which of the following could cause this inconsistency?

Explanation

A substantially smaller test database can fit more readily in memory or database caches, making reporting queries appear faster than they do against the much larger production dataset. If production does not achieve the same cache-hit behavior, it incurs additional I/O and reporting latency despite equivalent hardware and user counts. Performance testing should use production-representative data volumes and caching characteristics.

Learn more

CT-PT: Certified Tester Performance Testing

AstroTutor

Notes

Readiness

QuestionQ11

Community Discussion