Two factors are driving this growth, he continued. “First is the explosion in the quantity of data we store, especially unstructured/semi-structured data. Second is the combination of organic and institutional migration from traditional large-scale RDBMS deployments to more directed, lower-level technologies that more closely resemble the desired use cases.”
This momentum toward more focused, flexible databases shows no signs of letting up. Raj Rathee, head of product management at Exasol, expects “a continued uptick in growth of specialized databases as opposed to the one-fits-all-approach.” For example, he said, “We’re likely to see increased use of graph databases for representing and analyzing relationships, and NoSQL databases for storing IoT events. For analytics and data warehouses, in-memory and columnar databases—especially those that support MPP—will continue to be adopted for their unmatched performance and scalability for processing large data volumes.”
Jim Webber, chief scientist at Neo4j, sees graphs as the databases to watch over the year ahead. “They’ve reached the early mainstream now, and their impact for transactional and analytical applications is profound,” he said. “Early adopters are already profiting from graphs, and now the mainstream wants in. Graphs have been bubbling away under the surface for 10 years, but are at an inflection point now. The enterprise is no longer merely curious about graphs but hungry for them.”
Multi-model databases also represent “a new journey to handle the variety of data” according to an analysis published by Jiaheng Lu of the Department of Computer Science at the University of Helsinki. “The variety of data is one of the most challenging issues for the research and practice in data management systems.” Multi-model DBMSs build a single database platform to manage multi-model data, Lu said, noting, “Even though multi-model databases are a newly emerging area, in recent years, we have witnessed many database systems to embrace this category.”
Of course, this is a rapidly changing field, with varying business and technology circumstances, making it difficult to predict the path these data environments will take. Marc Caruso, chief architect at Syntax, sees a hybrid environment encompassing many types of approaches emerging. “Data migration is happening right now and at a large scale,” he said. “We are seeing continued migrations into open source relational database solutions, non-relational database solutions, PaaS-based database solutions, and a combination thereof.” The primary focus of these initiatives can be grouped under the heading of reducing operating costs, whether the initiatives are being undertaken to reduce hefty support contracts from major vendors, reduce headcount expense, or gain performance efficiencies by migrating to a more purpose-?built database solution, Caruso noted.
WEIGHING THE ADVANTAGES
Every database type has specific strengths for specific functions within today’s enterprises. Graphs, which are a type of NoSQL database, for example, can map very closely to business requirements. Graphs are expressive and can deal with the complexities, irregularities, and contradictions of modern business, said Neo4j’s Webber. “As such, users of graphs not only experience superb performance from their databases but find that the results they achieve are accurate and actionable. Moreover, graphs are the natural underlay for the best machine-learning systems.”