Workday and AWS let each side query the other without a copy

Workday, which makes software for staff data and money data, connects its Workday Data Cloud to Amazon Web Services (AWS). Each side can query the data of the other where it is, with no copy. Workday tells that this keeps the access rules and what the data means. The link serves the AWS tools for data and AI that customers already use. The announcement gives no test results.

How it works

Workday shows its data as Apache Iceberg tables, an open table format. AWS Glue Data Catalog records these tables. AWS tools such as Athena and Redshift then find and query them. The computers of the customer in AWS do the work.

Two directions

Workday agents can also query Iceberg tables in Amazon S3, a storage tool. Example: an agent for workforce planning gets revenue forecasts from AWS and gives advice on who to employ, for persons to examine. AWS Lake Formation sets which tables Workday can read. Workday security sets which users can see them.

What is not known

The announcement gives no test results, and speed and cost are not known. Workday tells that live queries operate in near real time. It gives no measured time.

Insight: Two firms share the permission layer for agents

On October 13, Workday told that its agents follow the limits of the person who uses them. On October 28, AWS told the same for its AWS Context layer. This connection joins the two. It seems to be a race to control the permission layer for agents, because the layer that holds the access rules controls the data flow. Monitor which layer controls access if the two disagree.