Abstract
In the current business environment, organizations face multiple challenges including outdated systems, newer technologies, and changing business processes. The airline industry is no exception. To manage passenger expectations and demand for seamless personalized services at each stage of the journey, a number of airlines are migrating current systems to different Passenger Service Systems (PSS). PSS migration programs are complex and critical as they impact most of the other applications and can potentially affect airline operations. One of the key factors to successful PSS migration is management of the test data and test environment at all stages of testing.
Ushering Fundamental Change
PSS governs critical airline systems such as airline reservation system, airline inventory system, and a Departure Control System (DCS).
PSS holds a wide range of product information, as well as important passenger data (such as journey history and customer preferences) that flows across applications. Changes in business strategy and obsolescence of existing systems necessitate the need for PSS migration. PSS migration involves business change, IT system change, and data migration, all of which trigger the need to train the business team. Changes in the airline industry are driving both large airlines and low cost carriers to consider migration to PSS. Cathay Pacific Airways and Singapore Airlines migrated from their in-house Mainframe systems to Amadeus; the LAN Group and Etihad left Amadeus for Sabre; Virgin Australia moved from Amadeus and Navitaire to Sabre.
PSS migration entails a fundamental change in the way airlines operate. Its successful implementation requires Test Data Management (TDM). TDM ensures that parallel testing activities can take place without impacting each other. This is particularly important for migration projects where it is critical that migrated data is not being modified by other processes. Parallel modifications may potentially result in incorrect test results and require additional effort in defect analysis.
Figure : Test Data Insights
This paper explores an effective approach to test data management for smooth PSS migration that reduces the effort spent in test data creation, maintenance, and idle wait time. It proposes that test data managers can ensure that test data is available to all without overlap and invalidation.
Containing Chaos: The Challenge of Migrating Data
The complexity of PSS migration has made TDM even more challenging. It is, therefore, crucial to maintain secure integrated data over repeated test cycles and other non-production environments. Ineffective governance policies results in the following challenges:
Limited availability of environment while multiple teams are accessing the same data There are only two to three environments available to perform development, testing, and training. Among these, one environment is designated for production. PSS data is accessed by a number of teams simultaneously viz. testing team, business team, PSS owner, and interfacing application owners. This means that there will be a number of user groups performing different activities in each environment. An integrated test environment poses the risk of test data overstepping, making the data invalid for other users. Aggressive delivery plan does not allow time delay or wastage Data mining, data preparation, and other data-related tasks consume around 40-50% of development lifecycle time. This means that the user groups must follow data management guidelines to ensure that there is no chaos or misconception when using data, especially migrated data in each environment. Specific data may be required for validating application functionality. User groups must also have a clear idea of their data requirements ensuring data variety, test coverage, and scheduled delivery. Non-production environments have space and performance constraints With full dump of production data not advisable for use due to storage constraints, a subset of production data is migrated. Given the large volume of data flowing across applications and interacting with the PSS and numerous other transactions generated from different use groups, performance is of vital importance. Judicial use of space necessitates liaising with different user groups and selection of correct subset of production data—ensuring full coverage of test cases and fulfillment of user groups requirements. All environments reside in a multi-hosted client environment These environments have maintenance activities occurring at scheduled times and user groups must work within this maintenance schedule. Any transactional or new data created by user will not be stored for use in later test cycles; hence test schedules and refresh cycles need to be properly aligned.TDM protects the usage of migrated data, ensuring that all user groups working in each environment can access both migrated and new data that suits their work scenarios and perform activities without impacting data of other user groups. Absence of TDM would result in chaos between the teams using test data indiscriminately, invalidating the test data created for other purposes. It would also result in delays arising from insufficient test data and conflicting test data requirements.
Test Data Management to the Rescue
Effective Test Data Management, a coordinated and easy-to-manage process, improves testing effectiveness and reduces time and cost of providing high quality, safe test data for smooth PSS migration. Our TDM framework depicts different components of test data management in sequential manner. The objective is to cater to the challenges user groups face and define best practices of test data management.
Figure: Test Data Management Framework
The PSS owner performs the following tasks to counter the challenges faced in test data preparation:
Data extraction Data subsetting and masking Data loading in different environments Configuration management Quality audits and data refreshTo ensure coordination and communication, organizations need a Test Data Manager who is responsible for:
Availability of test data for all test phases Coordinating with the Environment Test Manager and ensuring that the test data is available for testing Actively managing test data requests for applications and setting up data sets as per the master test strategy to ensure there are no delays in testing Ensuring that all user groups in each environment can access the required data and perform activities without impacting other user groups’ data.We also leverage on Test Data Management Governance approach that describes actions for effective test data management in PSS projects.
Different users perform different tasks in the test data management framework.
See the Bigger Picture
Several benefits are envisaged by following a streamlined process where all requests are funneled through the TDM team:
Reduced effort involved in the overall TDM process due to workflow automation, data request templates, and process improvement No rework and data overstepping with reservation and distribution of test data Reduced efforts with reusable tools Reduced time-to-market with no delays in test data delivery Increased defect detection efficiency due to availability of correct, timely data Improved productivity and overall efficiency Higher customer satisfaction and growth.
Streamlining Data Management
Effective test data management is crucial to ensure success of PSS migration projects. Dedicated test data management teams managing the test data setup and administration through appropriate TDM frameworks would help ensure accurate test data is provided in the test environment as per user requirements. The challenges in test data management can be addressed through implementation of a workflow-based request provisioning system for timely test data request, status and provisioning, tools for test data generation, effective coordination with user groups, improved defect detection, and increased customer satisfaction.
Tools providing a holistic solution to test data management like CA Test Data Manager can be leveraged to automate the test data management process. Data acquisition, subsetting, masking, test data generation, test data matching and coverage are some of the vital features of this tool toward improving the testing efficiency.
References
http://www.ca.com
http://www.sita.aero
http://www.amadeus.com