Page 1 of 4 next >>

Next-Generation Databases Provide an Embarrassment of Riches for Data Managers

If there has been one characteristic that has been emblematic of data management in the early 2020s, it would be the incredible variety of appoaches. Not too long ago, the data management space was dominated by either relational database management systems (RDBMSs) at one end or desktop-based databases on the other. Now, there is a database for every type of function—an embarrassment of riches for data managers. Cloud databases, multi-model databases, time series databases, graph and document databases, and a range of additional NoSQL systems all offer capabilities to address business situations across digital and data-driven environments. The question is: What is the best database environment for the purpose at hand?

The move to next-generation databases is driven by their ability to help companies achieve competitiveness and reach customers faster and more efficiently. These new breeds of systems can be a force for business transformation—whether it is generating new sources of revenue, enhancing customer experience, or producing data-driven insights that improve how organizations interact with customers.

“Advances in web technology, social networking, mobile devices, and Internet of Things have resulted in the sudden explosion of structured, semi-structured, and unstructured data generated by global-scope applications,” according to a published analysis by Ali Davoudian of Carleton University. Due to their inflexibility and the cost and complexity of data transformation and migration, traditional RDBMSs cannot meet many of today’s digital requirements alone, Davoudian and his colleagues stated. “Such applications have a variety of requirements from database systems, including horizontal scalability to linearly adapt to the massive amounts of data and the increasing rate of query processing by making use of additional resources, high availability and fault tolerance to respond to client requests, even in the case of hardware or software failure or upgrade events, transaction reliability to support strongly consistent data, and database schema maintainability to reduce the cost of schema evolution.”

NoSQL systems, for example, “are used not as a revolutionary replacement for the relational database systems but as a remedy for certain types of distributed applications involved with a massive amount of data that need to be highly scalable and available,” said Davoudian.


It isn’t just the NoSQL systems that are gaining attention—open source databases have also increasingly been implemented as solutions. “We’re seeing a huge growth in technologies like Apache Cassandra and Apache Hbase,” said Ken LaPorte, manager of the data infrastructure engineering team at Bloomberg. “We are also starting to leverage sharded database technologies, like Citus and Vitess, which build on our existing strong foundation of traditional PostgreSQL and MySQL deployments,” said LaPorte. “We’ve had numerous tenants successfully migrate their use cases to these sharded database technologies.”

Page 1 of 4 next >>


Subscribe to Big Data Quarterly E-Edition