AWS wants to offer customers better value for money within Amazon Redshift Serverless. AI must ensure that variable workloads are always optimally supported.

Amazon wants to optimize its Redshift serverless offering and is relying on AI to do so. The cloud specialist himself announced this at its Re:Invent event in Las Vegas. Redshift offers a cloud data warehouse that can process petabytes of data thanks to massive parallel processing capacity. The serverless version ensures that enough hardware is always allocated to support the requested workloads in Redshift through automatic scaling up and down. The more efficiently this is done, the better the performance and the more correct the end customer’s invoice.

More accurate scaling

Redshift already scales based on the number of queries. AWS is now using AI to make scalability more accurate. For example, query complexity can vary, as can the amount of data in Redshift, drastically changing the intensity of workloads. AWS gives the example of an organization that normally uses the service in a fairly predictable manner, but due to new regulations it suddenly has to process and analyze a lot more data. To ensure consistency in such a case, AI scales resources for data ingestion, new queries, pre-computations for quick access, data organization, and more.

The AI ​​also works proactively. For example, Redshift may require fewer resources during the day because administrators work primarily in dashboards. Later, the system needs to process complex queries, which requires additional horsepower. Customers can specify for themselves what price-performance ratio they are looking for. Because AI ensures that the infrastructure is precisely tailored to the workload, the result improves and you theoretically get more horsepower for the same money.