CERT
 
Research Staff Biographies CMU Heinz School CMU School of Computer Science CERT Statistics US-CERT CyLab
 

STAR*Lab System Component Composition

Engineering Automation for Understanding System Behavior

Principal Investigators: Kirk Sayre

Problem Addressed

Modern systems are characterized by large-scale heterogeneous networks with many components that must be correctly integrated to achieve mission objectives. It is often the case that the components are complex systems in their own right and must be dynamically integrated to provide end-to-end capabilities. System integration today is a complex, labor-intensive process that can require substantial effort for large systems. Automation support for behavior analysis of component compositions could help reduce the time and effort required to achieve operational capabilities [1].

Research Approach

This potential project will define the extent to which component compositions can be automatically calculated. Automation support for determining composite behavior of components architected into systems could enable fast and reliable understanding and development. Composition computation must generate mathematically correct abstractions of behavior at any level and help scale up the reliable unit of construction for systems. Because behavior calculation is essentially a compositional task, function extraction is the key underlying technology for component composition. FX produces behavior databases of individual programs; the databases themselves can be composed to reveal the composite behavior of the programs when combined into systems.

Expected Benefits

Automated derivation of the net effect of program compositions can reveal combined functionality, illuminate mismatches, facilitate analysis of design alternatives, and support evaluation of commercial off-the-shelf products. This approach can also guide rapid and reliable refactoring of components and systems in responding to new system requirements.

2007 Accomplishments

Research and development carried out in the FX project has direct applicability to automated composition of components.

2008 Plans

A key step toward creation of an automated composition capability is extension of FX technology to create a proof-of-concept prototype. Sponsors are welcome to join in this effort.

Reference

[1] Feiler, Peter; Goodenough, John; Linger, Richard; Longstaff, Tom; Kazman, Rick; Klein, Mark; Northrop, Linda; Wallnau, Kurt; Gabriel, Richard; Schmidt, Doug; & Sullivan, Kevin. Ultra-Large-Scale Systems: The Software Challenge of the Future. Pittsburgh, PA: Software Engineering Institute, Carnegie Mellon University, June, 2006. http://www.sei.cmu.edu/uls.

Disclaimers and copyright information

Last updated May 9, 2007