Using khazana to support distributed application development

One of the most important services required by most distributed applications is some form of shared data management, e.g., a directory service manages shared directory entries while groupware manages shared documents. Each such application currently must implement its own data management mechanisms, because existing runtime systems are not flexible enough to support all distributed applications efficiently. For example, groupware can be efficiently supported by a distributed object system, while a distributed database would prefer a more low-level storage abstraction. The goal of Khazana is to provide programmer's with configurable components that support the data management services required by a wide variety of distributed applications, including: consistent caching, automated replication and migration of data, persistence, access control, and fault tolerance. It does so via a carefully designed set of interfaces that supports a hierarchy of data abstractions, ranging from flat data to C++/Java objects, and that give programmers a great of control over how their data is managed. To demonstrate the effectiveness of our design, we report on our experience porting three applications to Khazana: a distributed file system, a distributed directory service, and a shared whiteboard.