Insights

AIMS (Airline Information Management System) software is among the most widely-used packages across the aviation industry, covering everything from scheduling crew rosters to assigning aircraft to different regional sectors. It is remarkably comprehensive, highly robust and very well-respected, which is why 180 airlines use it across Europe, the US, the Middle East and India. 

AIMS is so well-established that many airlines rely on it for their day-to-day operational efficiency. As one airline put it, if AIMS went down the company could only continue to operate for four hours. 

It’s why the testing of updates and new releases is so important for the airlines that depend on AIMS. Teams face the challenge of having to test every new code version of the software before it is introduced. If they get it wrong, fall behind, or are insufficiently thorough, it could have severe effects on operations and processes. 

The problem is that running manual testing on AIMS software is difficult – necessitating the testing of how every user journey or application works. There are many manual processes – and the tool is slow, taking protracted time to undertake each individual step.

The result is airlines conducting such tests need four or five people manually testing 2,000-3,000 cases every couple of weeks,  which is obviously time-consuming. Every six months 4,000 user journeys are executed ahead of major releases of the software. In the case of the airline we worked with, it was devoting four testers to this task, taking four months. They had to accomplish this alongside testing of minor software releases every month. 

This is fairly typical. These processes not only eat up resources, they are also error-prone as a result of the ‘confirmation blindness’ that can afflict human beings conducting repetitive tests of this nature, no matter how hard we may try to fight it. Those errors cause higher costs and slow down the speed of change.

Automated testing proof-of-concept and implementation of Sikuli

For one of Europe’s major airlines, we have shown there is a better way. 

Because of our domain knowledge, we know how AIMS software works and were able to establish an automated testing approach. We were aware it would be challenging. The implementation at this airline is based on Citrix, which is effectively a virtual machine but is configured on the IT management side, so it can run from anywhere while being very secure.

However, testing it is incredibly difficult. Being objectless software, not many tools are available on the market to test such an application. Nevertheless, our understanding of Citrix meant we found a solution.

The tool we employed is called Sikuli which uses image recognition to drive the automated testing process. We completed a proof-of-concept in just two days, after which Sikuli provided an automated view of the airline’s testing journeys. Automation slashed the time it takes to complete a testing journey from a couple of minutes to approximately 11 seconds. This is a significant time-saving when scaled to 4,000 user journeys every six months, as has been the requirement. 

Broader gains for all airlines using AIMS

What do we learn from this? It tells us that whilst AIMS remains a vital tool for airlines, there are numerous potential benefits to tap into by scaling up such an approach, centred around cost, quality and agility.

An automated full test suite could be set up to run out-of-hours, dispensing with the requirement for someone to be manually at their laptop. Not only is the actual user journey shorter, it can also run continuously at any time. 

By automating tests, the major airline has saved significant time and expense. Testers are able to focus more on exploratory testing rather than the drudgery of manual testing. And we have also effectively removed the scope for human error caused by confirmation bias, which can cause such unnecessary and costly problems.  

This is not an esoteric solution. Sikuli is user-friendly, easy to pick up and simple to understand. The goal of automation should wherever, possible, to be as reusable as possible. It may take groundwork at the start, but the potential to expand into different services, with little effort, is one of the many benefits we see rolling out from this approach.