watsonx.data

 View Only

Test design techniques and why is it relevant even now

By T K Narayanan posted 24 days ago

  

Introduction

With the ever changing, fast paced IT industry where the product/project deliverables have become more agile and with more emphasis on exploratory testing, one begins to wonder whether Test design techniques still hold any importance.

Test design technique

Before we delve into the topic, two of the standard test design techniques in general recommended by ISTQB (International Software Testing Qualifications Board) are:-

·      Boundary value analysis (BVA)

·      Equivalence class partitioning

As the terms suggest, BVA is testing around boundary values and Equivalence class testing is achieved by creating multiple equivalent parts for your tests.

Real time scenario

In this blog we will see a real time scenario that our team faced while testing Watsonx.data and how implementing BVA helped us in finding a crucial issue that if left unfixed could have possibly resulted in huge costs for the customers.

We have an option to bring in Spark Analytical engine into watsonx.data and the user can define the no:of nodes that is required as part of infrastructure. For the node definition, a user can enter values anywhere between 1 to 25. So from testing perspective our team decided to enter three boundary values (0,1,2) and (24,25,26).

After testing the positive scenarios, the team decided to enter 26 (in this case a number greater than 26 was also accepted) and the engine was provisioned. This turned out to be defect as team was also enter any value above 26 as no validation was implemented to restrict the number of nodes to 25. The team ended up provisioning a spark engine with a node as large as 90 and this resulted in a spike in cost for the user.

Since the testing was done in test environments and since we were able to catch this issue early it resulted in a major defect which could have resulted a huge cost of the end customer which could have been detrimental in deterring IBM brand.

On a concluding note, we were able to fix a major issue with direct cost implication to the customer by following the basic principles of test design technique by following “Boundary value analysis” .


#watsonx.data

0 comments
8 views

Permalink