Finding the suitable strategy for data migration

There are many reasons why a company is confronted with the challenge of a data migration. The most common reason is probably the replacement of an old system. In our blog post “Challenges of data migration“, we already discussed general questions that come along with a data migration. But what does it look like with the suitable strategy? What strategies are there and what one is the right one for your company and migration?

The strategies at a glance

What migration strategies are available is quickly clear. Some of you may have heard of the terms “chicken little” or “cold turkey” as general replacement strategies. There are two different approaches specifically on the level of data migration strategies: the iterative and the big bang data migration.

Both approaches offer different advantages and disadvantages. The right selection is a fundamental decision in every data migration project. The expenses for the project, the start date for productive data migration and the entire duration of the project are connected with this decision.

Big bang strategy:

With the big bang data migration, all data objects from the source system or the source systems including all forms and types are migrated on a reporting date or long weekend.

Iterative strategy:

The approach of the iterative data migration strategy is to divide the data objects from the source system or source systems into migration tranches, which are then migrated after each other in an iterative manner – meaning gradually.

A comparison of the strategies

The advantage of the big bang strategy is that there is only one migration reporting date and therefore only one migration balance. In contrast to this, the iterative strategy has multiple reporting dates and balances that have to be created individually. With the iterative approach, there is a transition period in which two systems must run parallel to each other, because a part of the data always only exists in the old system until the last tranche.

With the big bang data migration, in contrast, the source system is shut down on the reporting date and operation is completely converted to the target system. Furthermore, only a single migration business plan must be created and communicated only once with the responsible authorities (for example, BaFin).

Migrating the entire inventory of data objects with all its diversity on a single reporting date, however, also has disadvantages. Preparing the target system for big bang migration takes a long time. For example, all tariffs must be created in the insurance industry. This process is very time-intensive and, compared to the iterative strategy, leads to a later date of productive migration.

Moreover, a single migration date generates considerable peak loads for the employees and leads to the entire risk regarding the success of productive migration being concentrated on a single day or weekend. This increases the pressure on productive migration. With the iterative migration strategy, this pressure is divided through several migration dates and therefore reduced.

Another advantage of iterative data migration is an early start for the initial productive run. After definition of the individual tranches, they can be processed gradually. This includes all activities in the source system to provide data objects, to implement the migration process as well as to prepare the target system for the migration of every single tranche.

With several productive migration dates, a learning process is also created, which does not exist with the big bang approach. All departments that are a part of the data migration perform the process several times under the pressure of productive migration and thereby increase their experience and confidence. If you order the iterative steps according to the complexity of the data objects and start with the ones you believe are simpler, this approach leads to a well-trained team that is optimally prepared for the really demanding parts of data migration. This saves costs and nerves and leads to a higher success rate during productive runs.

The selection of the data migration strategy for individual projects should definitely be discussed early on. In addition to the pure comparison of advantages and disadvantages of both strategies, special requirements and stakeholders play a big role in the decision – depending on the industry and company. That is why it is important to get a good overview of the strategies as early on as possible. Because that is the foundation for the right selection of the optimal strategy for your project.

More information about the topic of data migration and our MIGSuite product, the software solution for professional, complex data migrations, can be found here.