In Amazon Redshift, what is the difference between the serverless and provisioned deployment models?

Amazon Redshift offers two deployment models: provisioned clusters and serverless. In the provisioned model, you launch a cluster with a fixed number of nodes; you pay for the cluster’s capacity regardless of usage, and you manage scaling, maintenance, and backups. Serverless, on the other hand, abstracts the underlying infrastructure. You simply create a workgroup and run queries; Redshift automatically provisions compute resources, scales them based on workload, and you pay only for the queries you run. Serverless eliminates cluster management tasks, provides on-demand scaling, and is ideal for unpredictable or bursty workloads. However, it may have higher per‑query costs and less control over the underlying hardware compared to provisioned clusters.