You can see some of Colin Watson's comments on the talk in his blog post. Also, an interesting thing (for me, at least) was that scanner benchmarking wasn't originally an intended use case for ThreadFix and its scan normalization and merging capabilities. But as it turns out those capabilities can be pretty useful for scanner evaluation. Bonus!Web applications pose significant risks for organizations. The selection of an appropriate scanning product or service can be challenging because every organization develops their web applications differently and decisions made by developers can cause wide swings in the value of different scanning technologies. To make a solid, informed decision, organizations need to create development team- and organization-specific benchmarks for the effectiveness of potential scanning technologies. This involves creating a comprehensive model of false positives, false negatives and other factors prior to mandating analysis technologies and making decisions about application risk management. This presentation provides a model for evaluating application analysis technologies, introduces an open source tool for benchmarking and comparing tool effectiveness, and outlines a process for making organization-specific decisions about analysis technology selection.
Slides for that talk can be found online here:Attaching web applications to databases as “sa” or “root” might be easy but it is also a horrible idea. This presentation provides a methodology as well as tools to create fine-grained database user permissions based on application-specific requirements. The negative impact of permissive database user account permissions is demonstrated alongside the potential benefits of constrained database user access. Tools for the automated creation of security-role-specific MySQL user permission policies will be demonstrated and these will be used as a model for making “least privilege” database accounts a standard practice in web application deployment.