There is no doubt that, historically, data migration projects have been prone to time and cost overruns and, in some cases, outright failures. However, that doesn't need to be the case. Experience over time has shown that there are actually three main risk factors associated with data migration:
- understanding your existing data adequately
- ensuring that data relationships don't break during the migration process
- assuring a smooth handover process at the completion of the project
Take appropriate steps at each of these points and you take the risk out of the migration process.
In this paper we will be focusing on the second and third of these points. However, that is not to downgrade the importance of understanding your data (which is technically known as data profiling). Data profiling is essential, not only to identify the scope of the data migration efforts that face you but also as a pre-cursor to ensuring that you don't break data relationships. After all, you have to know about these relationships (which is part of the role of data profiling) before you can maintain them. We should also say that we consider the use of appropriate data profiling tools to be essential: without them you have to profile the data manually, which is incredibly resource intensive. As a result, profiling is invariably inadequate and immediately introduces unacceptable risk into the migration process.
We will consider the issues surrounding the maintenance of relationships on the one hand and the cutover process on the other, with the aim of demonstrating how data migration processes can be de-risked. An outline is also provided of Celona's technology platform as a purpose-built product that meets the necessary criteria for a safe, flexible and reliable approach to data migration with respect to these two aspects of risk.