HDCCSR: Scalable Dependability in Componentized Software via Self-Stabilization

Ohio State University Research Foundation -Do Not Use, Columbus OH

Investigators

Abstract

---------------------- Scalable Dependability in Componentized Software via Self-Stabilization PI: Anish Arora, Professor, CIS, The Ohio State University This proposal addresses the problem of unanticipated software faults, which inevitably occur in large-scale software systems, including those used by the United States space program for missions such as the International Space Station, by developing tools and frameworks for self-stabilization, i.e., for restoring system behavior if it deviates from its intended one by detecting and correcting state predicates. Extant techniques for designing/composing stabilizing systems do not scale well. The proposed research addresses complexity as follows. - A component is typically simpler than the system as a whole. Hence stabilization will be added per component and then they will be composed into a stabilizing system, taking into account issues of dependency and interference. - A component's specification is usually simpler than the component itself. hence stabilization will be added per component with respect to only the component specification, The proposed research will demonstrate that the software tools and frameworks resulting from these ideas measurably and predictably improve dependability in a NASA system. The techniques and tools will provide for the first time the ability to make large systems tolerant to unanticipated faults. Besides space craft, they can also be used with critical systems such as aircraft control, medical instruments, the telephone network, etc.