Stress Testing

To Trace Stress testing is not just to determine stress point (when the system goes down or crashes because of increase in load) but also to determine the factors affecting the stress point. It also to determine the performance behavior of the system before reaching the stress limit.

TRACE APPROACH

TRACE provides an integrated complete end-to-end stress testing solution to its customers with its defined methodology and in-house testing tools. A well-defined end to end metric based process with proper entry and exit criteria focusing on the entire aspects of stress testing is followed at TRACE. The approach followed at Trace is based on determining the system behavior before reaching the stress point. This is done with a continuous iterative testing by increasing load. At each level all the performance attributes have been monitored and analyzed.

Attributes considered for Stress Testing

Following are the only few attributes out of many that considered during Stress testing

Response Time
Throughput
Screen transition
Time {Session time, reboot time, printing time, transaction time, task execution time}
Hits per second, Request per seconds, Transaction per seconds
Performance measurement with number of users.
Performance measurement under load.
CPU usage
Memory usage {Memory leakages, thread leakage}
Bottlenecks {Memory, cache, process, processor, disk and network}
Network Usage {Bytes, packets, segments, frames received and sent per sec, Bytes Total/sec, Current Bandwidth Connection Failures, Connections Active, failures at network interface lever and protocol level}
Web server {request and response per seconds, services succeeded and failed, server problem if any}.