The overall goal of DeeDS is to develop and evaluate an architecture for distributed active real-time database systems. Our strategy to handle the inherent complexity is to treat each function as orthogonal and including only the minimum required functionality.

This overall goal is further subdivided into i) restricting the behavior of the reactive mechanisms, and using contingency plans, in order to guarantee predictability; ii) introducing a deadline and value-driven dynamid heuristic scheduling algorithm, dynamically guaranteeing hard deadlines in the presence of soft deadline transactions and during transient overloads; iii) designing a loosely coupled predictable general purpose composite event monitoring facility, which, in particular, can service the reactive components in DeeDS; and iv) designing a replication and concurrency control mechanism that supports predictable and efficient database access in the distributed environment.

As a means for evaluating this architecture we are currently building a basic prototype (research vehicle). This prototype is being transferred to an industrial partner, and used in an evaluation of the architecture by using a pilot application.

Uppdaterad: 2008-11-27
Sidansvarig: DRTS-gruppen