In many cases, performance testing is conducted after the completion of some level of testing, such as integration testing, but we also consider the risks from a development perspective and propose a plan to conduct performance testing as soon as possible after an implementation that may cause performance problems is detected.
There are many cases in which the response time deteriorates as soon as a large amount of data is processed even after the integration test is completed, requiring major program modifications, or in the worst case, starting over from the design stage, which may result in unplanned costs.
We offer proposals to reduce costs through early detection.
Our development base allows us to set up large amounts of data, understand table diagrams, and prepare various queries for data preparation.
Implementing test automation does not always result in cost savings.
The implementation decision must be based on cost planning based on future projections of application life cycles and platform version upgrades.
For example, a SaaS that plans to deploy services over a long period of time will need to run the same scenario many times throughout the year, including regular program modifications, platform version upgrades, and periodic vulnerability assessments.
In such cases, automating regression testing can be expected save considerable costs.
We have a wealth of experience not only in verification, but also in development, operation, and maintenance.
We can make proposals that contribute to cost reduction for our customers, so please feel free to contact us.
|Application||Expense Reconciliation System|
|Process||Test planning / Test design / Test execution (system test, performance test, load test) / Results reporting|
|Environment||Windows Server 2016 / IIS 8.0 / Oracle 12c / Jmeter|
|Results||Performance test:In an environment with MAX data input for system requirements, Jmeter was used to execute all screen displays, actions such as buttons, form printing, and online batches, and the responses were measured.
Load test:Prepare the MAX number of simultaneous logins and accesses for the system requirements, and run scenarios on Jmeter to measure response and server CPU and memory utilization.
Result:Many response failures and out of memory failures were detected and evaluated by the customer.