Data migration risks

Even though indispensable, data migration is a common need in today's organizations which rely on computer systems. Every upgrade and every update indicate a question if there's a must to transform data to make it suit new conditions. Unfortunately, the answer often is 'yes'. Moving data between two systems (could be two different systems or just the two version of the same system) is generally called data migration and happens to be much more complicated and - therefore - risky than it seems at first sight.

Depending on chosen method, data migration requires temporary shutdown of all systems across the organization or a lot of time to accomplish. Both these options, even if they're not the only possibilities, are connected with risk. Risk in data migration depends on many factors including a few main groups. These are business risks, IT risks, and migration risks, followed with a bunch of others. Usually, the more attention is paid to data migration in its first stages, the lower is the risk later. In fact though, not all risk could be eliminated. And why is the risk so meaningful? Because every new-age company is dependent on data. And even a possibility that it could be cut off from data for an hour or two makes its managers shiver. Once data migration takes often at least a day or two, the results are easy to predict. However, well-prepared process makes no harm to data and lets computer systems work the way they're meant to.

Risk groups in data migration


Business risks

Profitability

One of the most common problems of today's business is meeting budget requirements. Data migration isn't different. Its costs can be divided into two main groups - direct costs and the indirect ones. The difference between them is simple. Direct costs can be more or less easily foreseen, while indirect costs are often a consequence of troubles which occur during the migration processes, therefore they cannot be predicted. These first costs are easy to consider in data migration budgets, while the second ones are the essence of data migration risk in case of profitability.

Reputation

The reputation issue applies also to the consequences of mistakes being a result of migration. Even though the minor ones can be easily erased and quickly forgotten, there always is a risk of mistakes that result in serious consequences difficult to hide from media for instance. An insurance agency to lose a part of its data would be a great material for the morning news.

Regulation

Customer data stored in enterprise computer systems has to be kept under special conditions which are indicated by the law. For instance, there is an extremely high level of security required in case of banks and financial data. But it doesn't mean that they could resign from data migration. In their cases, data migration is connected with another type of risk - the risk of failing to satisfy the law requirements. If this happens, company usually has to pay meaningful compensation fees.

IT risks

Data loss

Every data transfer is somehow connected with the risk of losing a part of data or information. Proper data migration requires data originated from one source be delivered to the final destination without any losses.

Application stability

Even though the migration seems to be really well-prepared and data transfer itself finishes with success, there is a risk of harm which new data can do to the target system. What it might be are different stability issues provoked for instance by a wrong structure of injected data.

Cut-over aborts

In addition to multiple data migration problems, there also is a cut-over abort issue which may be another consequence of errors in data transfer. What's more, cut-over aborts may happen even if data transfer seems to accomplish properly.

Extended downtime

Data migration is not only connected with the risk of not meeting budget requirements but also time requirements. It rarely happens that migration processes take shorter than expected but delays aren't the exception. Thereupon, every company going through data migration should be aware of the risk of extended downtime and, in a consequence, being not available to work as usually for a longer time.

Budget overruns

No matter how well data migration is prepared, it almost is impossible to predict all the issues, thereupon company managers should be aware of possible budget overruns. It's not a must but often predicted budget and real-need budget differ at a few points.

Delays

Preparing a comprehensive schedule of data migration definitely is a good practice, but even the most detailed schedule couldn't consider unpredicted delays which are extremely difficult to avoid. In practice, almost every data migration ends with more or less noticeable delay.

Data migration risks

Completeness risk

The more data migration is complicated and the more business objects it consists, the higher is the risk of incompleteness. It applies mostly to those additional objects which might be not the most important to the organization, but the role of mistaking them would also be noticeable.

Semantic risks

Sometimes it's just a matter of sign to completely change the meaning of expression. It might be data format, currency name, any unit. It also is one of data migration risks - settings of two applications, between which data is transferred, might be illusorily the same, but - in fact - differ in details. Forgetting about it may lead to complete change of sense of migrated data.

Corruption risks

The risk of data corruption in case of migration processes applies to the difference between data models of origin and target applications. It sometimes happens that the possible problem isn't discovered early enough and data transfer actually finishes successfully. Unfortunately, it might only be an illusion. Even if this wrong-model data finally manages to be loaded into the new application, the application itself might not work properly. Or - even better - might not work at all.

Stability risks

One of many reasons why organization managers decide on data migration is moving from instable applications to more reliable ones. Unfortunately, migration not always leads to improvement - it's one of many risks. Problems with target applications stability might be a result of thousands of reasons, including data coding issues on the top of the list.

Execution time risks

It isn't easy to predict how long data migration is going to take, nonetheless a lot of people do serious mistakes when trying to. At first, it isn't possible to scale out the time needed for transferring 1% of data to 100% and consider it certain. That's the most frequent cause of problems with data migration execution time.

Orchestration risks

Regardless of the size of the company that owns the system which data migration has to be done within, migration never is a single process. It always is a set of programs which have to be placed in the predetermined order. Why is it so important? Because often one process comes out from another, one prepares a base for another, so on and so forth. Therefore, skipping one process might result in failure of the following ones.

Dimensioning and interference risks

There also are many infrastructure risks which apply to the technical aspects of data migration and, mostly, the target application. What else can be distinguished are dimensioning risks and interference risks, the last ones connected with the fact that more people than required to migration are still connected to the system. What they usually are connected with are regular users who haven't been disconnected with the system by accident.

Target application risks

Finally, there also are the risks which aren't connected with the migration processes themselves, but with the target application. What might be wrong? In many cases problems may be a matter of new application's restrictions or its incompatibility with migration programs.



Nonetheless, the ones listed above still are only the risks which might (and should) be minimized. There have already been thousands of migrations finished with success so that really is nothing to worry about but to prepare properly, following the ideas and data migration best practices.