New serverless-computing platforms such as Google Cloud Functions, AWS Lambda, and OpenLambda provide RPC handling to cloud customers. The RPC handlers are inherently ephemeral; these services do not guarantee the persistence of state between invocations, so serverless-computing must be used in conjunction with a cloud database. We propose building an RPC-aware storage engine that runs alongside these serverless-computing platforms. There are several opportunities for tight integration: (1) client RPC calls introduce new, meaningful consistency boundaries; (2) RPC schedulers aware of database replicas can make better locality decisions; and (3) coordinating handler pausing with database change batches would improve performance. Our database will coordinate with handler workers via gRPC calls.