Consider a web application that is deployed to a single cluster. After configuring the application to use in-memory session replication, there is a noticeable drop in performance, possibly due to the large amount of data being placed in HTTP sessions.

Which action improves performance while continuing to provide high availability for this application's sessions?