The Vera C. Rubin Observatory's Data Butler provides a way for science users to retrieve data without knowing where or how it is stored. In order to support the 10,000 science users in a hybrid cloud environment, we have to modify the Data Butler to use a client/server architecture such that we can share authentication and authorization controls with the Rubin Science Platform and more easily support standard tooling for scaling up backend services. In this paper we describe the changes being made to support this and some of the difficulties that being encountered.
