Nmedia - Fotolia

Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

DBA skills must evolve, morph to cope with big data technologies

Big data tools are changing data architectures in many companies. The effect on the skill sets required by database administrators may be moderate, but some new IT tricks are likely to be needed.

New technologies are shaking up the status quo in data processing. Despite the rush of schema-less NoSQL and Hadoop platforms and associated tools, the change may be moderate when it comes to database administration. But some new skills will likely come to the fore for database administrators (DBAs) as systems based on those technologies get deployed in companies.  And the days when the DBA was the sole "keeper of the database schema" seem to be waning.

"There clearly is a shift going on, but it's not as extreme from the DBA point of view," said Joe Caserta, president of Caserta Concepts LLC, a New York-based consulting and training services company that focuses on data warehousing and big data deployments. In traditional enterprise data shops, teams of DBAs are still busy keeping well-entrenched relational databases and data warehouses built around familiar SQL running smoothly.

Developers have gained added influence on data design, however, and Caserta said the changes will push some DBAs to obtain added skills.

Things change, things stay the same

For example, the fact that developers can initiate Hadoop and NoSQL projects with little or no upfront schema represents a change in enterprise development methods. Even so, companies are still likely to create reference data models after the fact. And while such modeling may be done by enterprise architects or data architects, some DBAs also play a modeling role -- so they'll need to learn how NoSQL systems work, Caserta said.

"They'll be using different tools and different modeling strategies," he said. "We still need someone to come up with the models. We also need someone who knows how to administer these new databases."

Handling the latter task likely will require new training: A DBA who is certified on the Oracle database is not going to be able to automatically set up and design, say, a Cassandra NoSQL database. "The methods the typical DBA needs to know have to be relearned," Caserta said, adding that the ability to manage Hadoop clusters will also become a crucial skill for DBAs. In general, "they'll need to learn how to store data when they don't have a schema."

What price agility?

Craig Mullins, president and principal consultant of Mullins Consulting Inc. in Sugarland, Texas, said the NoSQL movement has precedents that will ease the learning process for some DBAs. "It's not as new as some people think," he said, noting, for example, the similarity between the mainframe-oriented VSAM file technology and newfangled key-value data stores.

There are some fundamental differences, though. The agility offered by NoSQL software comes at a price as data integrity becomes far more challenging to achieve. But for now, full integrity is taking a back seat to data flexibility in many companies' Web-based applications. For the DBA, the big challenge is to adjust as design and development styles shift.

"There are some DBAs who are more adaptable than others. We've already seen that play out in the last 20 years," said Mullins, who has more than 30 years of data management experience and is the author of Database Administration: The Complete Guide to DBA Practices and Procedures, first published in 2002 and updated in a second edition in 2012.

Mullins pointed out that, in many organizations, DBAs are already expected to work with several relational database systems. DBAs with those kinds of skills "should investigate the NoSQL options so they can be in the forefront of guiding the organization wisely when and if they need the NoSQL offerings," he said.

Looking forward, Mullins added, data schema definition may be minimized, but "the availability of systems and the understanding of how data is spread across nodes will become even more important."

Salad days for DBAs?

In fact, Sue Geuens, president of DAMA International, an association of data management professionals, thinks new data architectures provide opportunities for DBAs to expand their usefulness in organizations. "DBAs have been seen as the techno-geeks who sit in corners and nurse your database servers through all their ups and downs. And I think that DBAs are getting somewhat tired of that label," said Geuens, who is based in Johannesburg, South Africa, and works as head of data services at EPI-USE Systems Ltd., an SAP-focused software and services provider.

Geuens said we likely will see different types of DBAs in the future, with some being content to stay in a traditional technical and administrative role, while others will look at learning the new technologies and tools for managing big data. "We will see a slew of new job titles for DBAs, and we'll also see specialization even deeper than the ones we currently have around the specific relational database," she said.

As Geuens emphasized, adding new skills is nothing new to the DBA job, which already comprises a full menu of capabilities. Pertinent skills include modeling, performance management and basic administration, with specializations for each database brand adding to the job's complexity.

Such expertise can mean significant pay. TechTarget's 2014 IT Salary and Careers Survey, for example, showed database administrators among recent gainers with average total compensation of $115,630. That's up 22% year-over-year.

High salary or not, working with new technologies such as NoSQL databases does require you adjust your way of thinking, said Greg Novikov, a database specialist at New York-based insurer MetLife who spoke at the MongoDB Days 2014 event held in Boston in October. "But that's the reason we're getting paid the big bucks," he added -- with a wink to the audience.

Jack Vaughan is SearchDataManagement's news and site editor. Email him at [email protected], and follow us on Twitter: @sDataManagement.

Next Steps

Find out more about the role of the DBA

Get background on DBA staffing considerations

Learn about the types of DBAs

Discover three new ways to build a big data system


Dig Deeper on Database management system (DBMS) architecture, design and strategy