6 Major Mainframe Migration Project challenges and how to avoid them

Transformation, legacy transformation, migration, mainframe migration, legacy modernization, modernization, code conversion, code transformation, Asysco, Asysco migration technology, migration technology

Interview with migration veteran and Chief Project Officer Rob Lawrence

The mainframe is usually at the heart of the business. The mainframes that are still in use today mostly run business critical applications that are custom made and have been in use for many years. All easy to migrate or to replace applications have already been taken care of. What remains are the complex and critical legacy systems.

These migration projects have a huge impact on the business, come with many challenges, involve multiple teams and disciplines and require extensive testing. In this interview we talk with Rob Lawrence, Chief Project Officer of Asysco about the 6 biggest challenges that organizations face during a Mainframe Migration Project.

Challenge 1.       Availability of key resources

Rob: A migration project requires the experience and expertise of many Customer resources. Often these resources are already busy with business as usual daily operations which may conflict with the needs of the migration.

We come across this challenge quite often. The advice that we give varies per organization and per situation, but I have listed a few tips and recommendations that have proven to be very useful over time:

  • Evaluate team capacity in advance. Consider business-as-usual and other projects that will overlap.
  • Adding temporary resources may be an option, but be wary of diluting experience and knowledge in the migration project team.
  • Dedicate a Technical Lead to the migration project that is experienced in both the mainframe application and Company standards and procedures.
  • Be realistic about resource availability. Planning for an extended duration project at the beginning of a project is more favourable than falling behind and delaying.
  • Personal commitments, annual leave and unexpected events will inevitably occur during a long duration project. Allow sufficient contingency when making plans.

Challenge 2.       Estimation of the test effort

Rob: Testing forms a major component in a migration project and having an accurate assessment of the effort required is a precursor to success; however, this can be challenging to predict. With the extensive experience we have acquired over the 20 years that we have performed mainframe migrations at Asysco, we’ve learned some valuable lessons that we bring into every project, I have listed the most critical success factors below. 

  • Assign a dedicated Test Manager in advance of the start of the project.
  • Utilise the time between the initial code extraction and first delivery to develop a thorough test plan.
  • Develop usable and informative metrics to track progress on a daily or weekly basis.
  • Engage early with the business to identify and prioritise testing. Knowing, and having agreement on the scope allows for more accurate estimates of the duration.
  • Test execution is only one part of the test effort – factor in time for investigation of differences, regression testing and re-testing of resolved defects.
  • Expect the unexpected – even the best thought out test plans will need to adapt constantly during the project. Allow for threads of testing being blocked by complex defects and have alternative threads lined up.

Challenge 3.       Reference environment

Rob: During testing there will be many differences identified, though not all of these will be defects. Having a mainframe environment with an identical code base and data set is necessary to debug and investigate existing behaviour. Important things to consider are: 

  • Investigate the possibility of creating a dedicated reference environment on the current production, development or DR mainframe (consider available memory, disk and licensing restrictions).
  • Factor the need to execute several iterations of batches, data extractions and processor intensive jobs into usage calculations on metered mainframes.
  • If hardware resources or other constraints prohibit a dedicated environment, consider how BAU development and operations will impact the ability to reliably use what is available as a reference environment.

Challenge 4.       Testing infrastructure

Rob: Typically a test plan will call for multiple threads of testing to be executed in parallel, for example, on-line and batch. In this situation it will be necessary to have independent testing environments to prevent each thread interfering with the results of others. In order to make this process as efficient as possible we usually recommend the following:

  • The use of virtual machines makes this much less of an issue that would have traditionally been the case.
  • Ensure that the VM hosts have adequate capacity to support all planned, and some un-planned, environments.
  • Verify procurement procedures in the organisation, if it is not possible to provision VMs at short-notice is there a way of obtaining pre-approval.
  • Aside from VMs, it is likely that multiple back-ups of the database and mass storage files will be generated during testing. Confirm the availability of sufficient storage space to accommodate these.
  • The use of compression for back-up files is strongly recommended to keep space requirements to a workable level.
  • In most cases, test environments are transient and will be removed at the end of testing. Sometimes it is appropriate and possible to consider non-typical hosts, for example the to-be production environment.

Challenge 5.       Problem Management

Rob: A major factor for successful delivery of the project is the Project Manager. To be an effective Project Manager on a complex project requires an in-depth knowledge of the process and workings of the organisation and how to navigate the various challenges that will occur throughout the migration. Consider the following when selecting a Project Manager:

  • Recognise that a project of this nature, that will touch in some way almost every area of the business, is a complex undertaking that requires the dedication of a full-time project manager.
  • Strong communication skills are essential.
  • Wherever possible, select a Project Manager that is familiar with the organisation. Having the allocated Project Manager engaged at the discovery stages of the project will help develop their understanding of the intricacies and technical challenges ahead.
  • An automated migration is a unique project and one that a Customer Project Manager will be unlikely to have experience of. Being receptive to guidance and support from the Asysco Project Manager is more of a benefit than prior experience.

Challenge 6.       Test data 

Rob: Whilst some organisations have fully anonymised test data already available it is not unusual that either old production data, or a subset of, is used for internal development. During a migration project it will be necessary for Asysco personnel to work with the same data that is being used for testing, without which we would be unable to reproduce and investigate issues and defects. Depending on the sensitivity of the data there may be regulatory or organisational restrictions placed on how and where non-anonymised data can be used. In order to make the testing process as effective and smooth as possible and avoid conflicts during testing, the following considerations need to be made in advance:

  • There are options available to the project in order to accommodate restrictions on data, such as ensuring it always remains within the confines of the customer’s organisation.
  • Before developing a test plan, consider the sensitivity of the test data and any limitations that are in place. If approval is needed to satisfy the requirements seeking this early in the project will avoid potential delays later.
  • If anonymised data is to be used, consider how accurately this can reproduce production behaviour.
  • At some stage in the project it is highly recommended that testing is done using a full set of production data. Consider how issues identified in this phase and how defects may be translated and reproduced within the data Asysco has access to.

Thanks, Rob for giving so many great tips and recommendations. If you look back on the many migration projects that Asysco has done over the last 20 years and the 100% success rate, what is the secret for success?

Rob: Our experience and proven project methodology allows us to confidently plan and manage our projects.  We have a very experienced team at Asysco and with that comes the ‘know-how’.  Couple that with our open, honest and transparent culture, with a ‘no blame’ attitude and we can navigate confidently around issues as and when they arise. We believe in true partnerships where we align to a common goal for project success.

About Rob Lawrence 

Rob joined Asysco in 2009 with 20 years of experience in IT Change & Operations.
Most of that time was focused in the Finance & Banking sector. Since joining Asysco Rob has been responsible for the successful delivery of the global project portfolio, working all across the world and has been involved in over 40 successful  mainframe migrations. He spends most of his time in Europe and North America. Rob studied Computer Science at Filton Technical College and is a certified Project Manager. He is also a member of the Asysco Executive Management Team.