Perpetual Model Validation
Abstract
This research effort investigated fundamental techniques to provide perpetual model validation, where design-time models are validated continuously during runtime. The research considered two fronts: validation of the software model, and validation of the model of the system interacting with the physical world. Since the field of runtime verification has extensively focused on the challenge of runtime model validation in software using direct models, the approach employed here instead considered using indirect models of software execution, for example memory access patterns, to check for security intrusions. Additional research was performed to tackle the essential problem of model validation for systems which contain interactions with the physical world using hybrid automata models. Perpetual model validation will ensure that the actual behavior of the system conforms to the analysis model, raising the level of confidence that can be placed in the results due to formal nature of the analysis.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 2017
- Accession Number
- AD1028363
Entities
People
- Matthew L. Anderson
- Stanley Bak
- Steven Drager
Organizations
- Rome Laboratory