| United States Worldwide |
|
| United States Worldwide |
|
A Review of the Rationale and Architectures of PJama: A Durable, Flexible, Evolvable and Scalable Orthogonally Persistent Programming PlatformMick Jordan and Malcolm Atkinson Abstract: A primary goal of research into orthogonal persistence is to simplify significantly the construction, maintenance and operation of applications in order to save software costs, extend the range of applications and improve users' experiences. To test such claims we need relevant experiments. To mount such experiments requires an industrial-strength persistent programming platform. The PJama project is an attempt to build such a platform and initiate those experiments. We report our design decisions and their consequences evaluated by four years of experience. We have reached a range of platforms, demonstrated orthogonality and provided durability, schema evolution with instance reformatting, platform migration and recovery. The application programming interface is now close to minimal, while we support open systems through a resumable-programming model. Our architecture is flexible and supports a range of optimisations. Performance measurements and current applications attest to our progress, but it is still possible to identify major research questions, and the experiments to test the utility of orthogonal persistence are still in their early stages. Last revised Sep 29th 2000 Questions and comments to forest-info@sunlabs.com | ||||||||||||||||||||||