The text discusses how modern distributed databases like CockroachDB handle the demands of IoT and microservices applications by using multi-row SQL Data Manipulation Language (DML) and parallel processing, which significantly enhance throughput compared to traditional single-row DMLs. It highlights that CockroachDB, since its inception, supports multi-row DMLs, offering an order-of-magnitude improvement in performance by reducing network round-trips and resource consumption. By using examples such as a tabulation service, the text illustrates the efficiency of multi-row DML in processing multiple client-generated events simultaneously, overcoming throughput bottlenecks. It also explains that while increasing batch size generally improves throughput, there's a point of diminishing returns due to factors like network latency and available resources. The text further contrasts the challenges of over-parallelization in single-node databases with CockroachDB’s ability to automatically scale and handle concurrency without extensive planning, emphasizing its suitability for unpredictable workloads typical in IoT and microservices environments.