Companies are increasingly chasing zero-RPO and zero-RTO solutions to ensure that their applications remain online no matter what. And it’s easy to see why: according to 2020 research from Gartner the average cost of IT downtime is $5,600 per minute.

Internet of Things (IoT) and microservices-style applications need a database that can handle requirements such as fluctuating number of client connections, unpredictable workloads, and bursty throughputs. Traditional single-node databases handle these requirements by reducing latency to improve throughput. However, for modern distributed databases such as CockroachDB, the optimal approach to handle these requirements is to use multi-row SQL Data Manipulation Language (DML) and parallel processing. Multi-row DMLs provide an order-of-magnitude improvement in throughput performance as compared with equivalent single-row DMLs, which is why databases such as Oracle, MySQL, and Postgres widely support multi-row DMLs. CockroachDB has supported multi-row DMLs since the 1.0. This blog post discusses how to use multi-row DMLs, the performance benefits of multi-row DMLs over single-row DMLs, and the effects of compounding database and application parallelism in single-node vs. distributed databases.