When businesses are looking to deploy big data analytics solutions, one factor that must be considered is what database technologies they will be building their systems upon. In many cases, the attitude among IT pros may be to stick with what they know and persist with relational databases, which they will usually have a high level of familiarity and confidence in. But is this really the best solution for every scenario?

As the volume and variety of information gathered by businesses grows, many enterprises may find these solutions struggling under the strain, particularly as they have to incorporate more unstructured data sets into their activities.

Matt Allen, senior manager of product solutions at NoSQL database specialist MarkLogic, explained to eWeek: "When relational databases were first developed, data was seen as small, neat, structured and static, and that's the only way it could be stored. Today's data is anything but that."

As a result of this, businesses can no longer rely solely on one-size-fits-all relational database models if they wish to get the most out of their big data analytics initiatives, but instead consider newer solutions that are better equipped to handle current workloads.

One common problem with relational databases is they tend to be very rigid, with the information laid out in rows and columns. While this works well for traditional, structured data, it means they struggle to cope when businesses are gathering data from a wide variety of unstructured sources.

"Relational databases can be configured to accept data variety, but only with cumbersome changes that result in additional schema complexity, or in ways that do not get the full value out of the data," eWeek noted.

This also means it can be very difficult to adapt existing databases, whether to scale up to larger volumes, accept new data, or handle mixed workloads.

Depending on the system, eWeek stated it could take months or even years to set up data modelling solutions in these databases, and once these are in place, making changes after the fact can be a hugely time and resource-intensive process. In an environment where big data demands are changing all the time, this lack of agility could be a major barrier to the success of a project.

"Regardless of any re-engineering, relational databases are simply not designed for the scale at which today's businesses operate," eWeek stated. "Trying to make them scale typically results in buying more and more expensive hardware and handling periods of disruption while changes are made."