There are a variety of events that lead to a data migration to the cloud, from a merger or acquisition to simply looking for better performance from a software solution. The drastic increase in data produced for companies is making a data migration a more significant event. Choosing the right path is generally the first step to enjoying the lowered costs, increased flexibility, and improved performance that can be expected in a cloud data server.
There are three major approaches to data migration, and each carries its own benefits and drawbacks:
Big Bang Approach: This all-at-once method has some benefits to offer, including a reduced cost of the total migration, and the ability to face down the project with a single deadline. It may be somewhat painful, but will be limited in its timeline — at least in theory. You won’t need to run two systems at once or wait for months to see if you’ve achieved a positive return on investment.
The drawbacks are that a migration like this is generally completed over a single weekend, and if you run into any roadblocks, you may have a team show up on Monday morning without a working data system. This type of data migration is considered high risk and is best used in situations where there is a smaller set of data or a system that isn’t critical for daily operations. It also requires that the company isn’t running a system that needs to be functioning 24/7.
Parallel Approach: This method offers the lowest level of risk because the new system is run alongside the legacy system until all the bugs are worked out and the team can ensure that business processes will run properly. The advantages of this choice are that the business is not disrupted during the migration, and any issues can be fully addressed before the legacy system is turned off.
The main drawback of the parallel approach is that it can get costly running two systems alongside one another, and if it takes a long time to work out any problems, the company can really pile up costs.
Phased Approach: In this type of data migration, the data is shifted in segments, with phases designated by module, volume or another designation in the system. As each segment of data is transferred, bugs can be worked out, which could make it easier to train employees on a new system. The phased migration allows for celebration along the way as each segment is successfully migrated.
The drawbacks of the phased approach include its tendency to become expensive as the project is drawn out, but it may strike a nice balance between big bang and parallel migrations in terms of cost and risk. The enterprise can move one branch office at a time, or begin with less-critical data to be migrated to the cloud to reach a plan that balances risk, cost and speed.
To discuss the best possible approach for an upcoming data migration, contact us at eXemplify. We work with the leading data centers in the industry and can help you meet your unique data center needs.