CAREER: Contingent Static Analysis for Dynamically Composed Systems

University Of Texas At San Antonio, San Antonio TX

Investigators

Abstract

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5). Pervasive societal use of information systems has increased the need for software to be both reliable and extensible. One technique that can be used to improve software reliability is the application of static analyses to discover and validate that properties hold across all program executions. Unfortunately, additional challenges arise when applying static analyses to programs that can be dynamically extended, such as is the case, for example, with Firefox plugins, Microsoft Office Add-Ins, and Eclipse plugins. This project will develop techniques for contingent static analysis that will be able to more precisely and efficiently analyze dynamically extensible software systems. Specifically, this project will explore techniques for accommodating dynamic composition during interprocedural program analysis using a runtime mechanism to efficiently avoid making overly conservative approximations. Analysis results will be expressed as contingent properties of different program regions, where the validity of each contingent property depends explicitly on the validity of properties from other program regions. The runtime mechanism will then validate and invalidate these contingent properties as new code is dynamically loaded into the execution environment. This approach will be applied to develop contingent escape and information flow analyses.