programming languages and tools that may result in verifiably secure systems. He contributed to the development of the Common Vulnerability Scoring System (CVSS) v3.0, and is also active in developing vul-nerability disclosure policy and standards as a member of the FIRST Vulnerability Coordination SIG. Previously, he taught mathematics and computer science courses as an Ad-junct Instructor at several universities and tech schools in the Pittsburgh area. Garret received MS and BS degrees in both Applied Mathematics and Physics from the University of New Orleans.
3. Panel Summary
Marco first listed major challenges for mobile security re-search: (i) usability, (ii) security too strong and not usable is useless, (iii) even children use mobile apps, (iv) we need to make usable reports, (v) we need to eliminate false positives so that users will be confident in results and only true posi-tives are reported. With respect to the question of how to deal with those challenges, he believes we must produce easy-to-configure tools, which provide incremental and easy-to-use feedback. Also, he thinks we should combine program anal-ysis with machine learning, to eliminate false positives. Looking 5-10 years into the future, he thinks we will develop more secure mobile apps by delegating a lot of security checks and enforcement to the underlying OS and middle-ware, and more carefully ensuring that data will remain pri-vate.
Garrett identified major security research challenges as: (i) being able to identify what information that apps are us-ing, (ii) what is going on at the network level is not always clear (how data flows from wireless provider to cloud ven-dor), (iii) supply chain issues with radio/baseband hard-ware/firmware (we don't know which models of devices contain which parts, making it hard to track all vulnerable instances down and get them patched with updates) (iv) firmware tends to be proprietary, not open source, making it hard to test. To deal with these challenges, he thinks that (software and network) fuzzing and automated testbed use should become widely used by developers and security test-ers. Further, automated testing tools still require a lot of man-ual interaction and he thinks more should be automated. Cur-rently, he thinks most developers do not understand basic methods for coding secure apps. In