Many new database systems are emerging for specific purposes, replacing the era of the single-purpose database. Enterprises will have to choose between the ease of use of more general-use databases and the performance of the new databases over the next few years.
The era of the single-use database is over and we are seeing a plethora of new database systems built for specific tasks (distributed key value stores, NewSQL, graph databases, etc). Over the next few years there will be some shakeout in the industry as enterprises decide between the tradeoff of performance that the new databases bring versus the ease-of-use that more general use databases (MySQL, PostgreSQL) offer.
Of particular note is the emergence of streaming data, which requires data to be analyzed while still in the server’s memory. Here, Kafka and other technologies supplant the traditional Hadoop and Spark-based big data tools. Also, there is the ongoing challenge of fitting stateful database workloads into “stateless” cloud-native architectures.