As distributed systems span the globe, placing data near the point where the data are accessed is becoming important to improve client performance and to reduce network load. With the advent of the utility model in storage services, it is necessary for storage service companies to provide quality of service guarantees to meet customer needs. Adaptive data migration will be required by storage service providers to guarantee these service-level agreements. Storage customers will require the ability to transparently migrate data among different storage service providers based on, for example, pricing differences or QoS issues. Data migration is also useful for reducing resource contention (thus increasing scalability) by breaking up "hotspots" in the file system. Data migration plays a significant role in increasing the scalability and performance of distributed file systems. In this paper, we examine mechanisms for transparent, flexible data migration in the context of our experimental distributed file service, DiFFS. 11 Pages