Getting Past the Big Bang Myth to a Workable Data Migration Strategy
Steve Winkler | October 12, 2016
One of the fundamental decisions an organization must make about its data migration strategy is when to migrate information from existing systems into a modernized platform.
After all, data migration can be a fairly disruptive process, so there’s a strong appeal to getting it over with quickly, even if an outage must be imposed upon users and stakeholders. The “Big Bang” approach is a popular term that describes a strategy whose goal is to shorten and simplify the movement of data from the old to the new system. It’s the veritable “rip the band-aid off,” brute force, 800-pound gorilla of the data modernization world. Sometimes it’s the right approach, but only if the conditions are correct.
How Big Bang migrations work (in theory)
In an ideal world, an organization could shut down its systems at close of business on Friday, and move the data on its legacy system to the new platform in a matter of hours. Theoretically, they could flip the switch on Monday morning, and all the data would be in place on the new platform and ready for use.
If a data migration strategy could actually work this easily and quickly, it would greatly reduce outages, disruptions, and confusion for data users and other stakeholders. But in my experience, such results are rarely, if ever achieved. The reality is that such an approach can only work when an agency is moving relatively small amounts of data, or for a limited number of offices.
How Big Bang migrations work (in the real world)
For most organizations, the complexity and volume of their data — and the fact that stakeholders need uninterrupted access to it — mean that the Big Bang approach is not a workable strategy. Stakeholders are not incentivized to agree to long-term outages, and even when they do, Big Bang migrations can encounter unexpected data or other issues which quickly turn an acceptable outage window into an unacceptable one. Fortunately, there are alternatives that program managers can use to mitigate some of the risks associated with the Big Bang approach.
One option is to take a phased approach in your data migration strategy, rolling out or piloting the new system for a few offices at a time. In certain cases, such an approach could allow an agency to have specific groups of users literally log off the old system on a Friday and access the data on the new system on Monday with little or no interruption. In other words, it can deliver results similar to those promised by the Big Bang approach, but on a far smaller and more manageable scale. Outages are shorter, and a smaller portion of stakeholders are affected by any given migration window.
Another strategy is to use a data synchronization solution to give stakeholders greater flexibility in using both the new and old systems simultaneously — at least until the data migration process has been completed. This approach offers the potential for capturing data changes that are occurring in either the old or new system and applying them in the other. By deploying such technology effectively throughout the migration, users can continue to access and make changes to the information in the old system, with their changes also showing up in the new system (and vice versa).
Such a synchronization solution is not necessarily designed to stay in place forever. But in the short term, it can help program managers ensure that stakeholders can continue to see and use the information in the old system, as well as reflect any changes they make there in the new system as well (or vice versa). Some sort of synchronization is almost always needed when migrated users need to share information or participate in system transactions with legacy users.
It’s important to point out that synchronization approaches are not always the perfect solution, either. They can be complex, tricky to plan and deploy, and if not managed and implemented correctly, they too can result in data discrepancy and confusion among stakeholders. The cost and risk of synchronization approaches must be weighed against the operational and mission flexibility they bring.
Keep your eye on the finish line
The main takeaway is to be aware of the three approaches described here — Big Bang, phased, and synchronization-enabled — and to know that each has its strengths and limitations.
Ideally, an organization with complex data and multiple sites will consider each of these approaches and use the elements that best meet their needs. By creating and implementing a customized strategy, you can give your program managers greater flexibility in the way their data is migrated, and at the same time, give data users more options for seamlessly executing their missions during modernization efforts.