A growing fragmentation in database offerings also suggests a significant sea change in enterprise requirements. “Any time the market fragments, it’s an indicator that there is no general-purpose solution,” observes Kurt Dobbins, CEO of Deep Information Sciences, Inc. And this stems from the fact that “there is no single solution that can handle the demands of the modern enterprise,” he says. “Relational databases, whose core algorithms were created in the 1970s, have not kept pace with the need for enterprises to keep and index everything to support ad hoc queries, while maintaining transactional integrity for business operations and the back office.”
As a result, “a huge variety of databases and frameworks has been spawned attempting to address these specific areas of deficiency,” Dobbins continues. “Where ad hoc queries are needed, unstructured solutions have emerged, all under the umbrella of NoSQL. With NewSQL solutions, you don’t have to worry as much about managing some of the more difficult aspects of relational databases, such as ensuring transactional integrity, when you want to make improvements.” NewSQL, for example, makes it easier to move data to in-memory technology, he says.
Relational Databases Reconsidered
However, nobody expects the relational database providers to sit idly by while these changes sweep around them, Imhoff says. The large players are already adapting and extending their offerings to address many of the challenges of managing unstructured data. “Although the use of alternate database technologies will continue to increase, traditional databases are expected to be the most used database technology in the enterprise,” says Alex Attal, executive vice president of enterprise business services for Fujitsu North America. “Our expectation is that traditional database vendors will gray the lines between traditional and nontraditional databases, expanding their product features to include key aspects of alternate database technologies.”
Growing fragmentation in database offerings suggests a sea change in enterprise requirements.
Rajeev Nayar, associate vice president and head of Infosys’ big data practice, for one, points out that “it will not be very easy to move away from relational databases in the near future, as newer databases are yet to mature.” In addition, he says, “it is not easy to replace traditional databases owing to existing investments and related processes that have been put in place.” Rather, newer applications and innovations are the approaches for updating data environments. Still, “it’s important to understand that each database has its role based on the workload that needs to be processed. We will eventually see a happy coexistence between Hadoop, RDBMS, NoSQL and in-memory solutions,” he predicts.
Variety is the Spice?
Is all this variety and innovation a good thing? Some industry experts advise caution. Security becomes even more of an issue as data gets distributed across enterprises, in new formats. “Databases are becoming a target and many NoSQL systems come with minimal security features,” says Joe Santangelo, principal consultant with Axis Technology, LLC. At the same time, he adds, “data resources are getting more complex while providing an abundance of information. We must take care to guard these resources and prevent the data from being accidentally exposed or maliciously exploited.”