Adaptable distributed transaction systems
Abstract
There is a need for adaptable systems that can change as requirements change, both in the long and the short term. Adaptability can reduce software costs and improve reliability and performance. Adaptability includes design techniques that support enhancement, retargeting to new projects, restructuring during different phases of a mission, and changing algorithms as external conditions change. These techniques are particularly important for projects with stringent performance or reliability requirements. We model algorithmic adaptability, which supports techniques for switching between classes of schedulers in distributed transaction systems. RAID is an experimental system implemented to support experimentation in adaptability. We study adaptability features in RAID, including algorithmic adaptability, fault-tolerance, and implementation techniques for an adaptable server-based design. Experiments are described that quantify the costs and benefits of adaptability.
Degree
Ph.D.
Advisors
Bhargava, Purdue University.
Subject Area
Computer science
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.