Optimizing a Network Layer Moving Target Defense by Translating Software from Python to C
Hardman, Owen Russell
MetadataShow full item record
The security of powerful systems and large networks is often addressed through complex defenses. While these types of defenses offer increased security, they are resource intensive and therefore impractical to implement on many new classes of networked systems, such as mobile phones and small, embedded network infrastructure devices. To provide security for these systems, new defenses must be created that provide highly efficient security. The Moving Target IPv6 Defense (MT6D) is a network layer moving target defense that dynamically changes Internet Protocol version 6 (IPv6) addresses mid-session while still maintaining continuous communication. MT6D was originally written in Python language, but this implementation suffers from severe performance limitations. By translating MT6D from Python to C and taking advantage of operating system specific application programming interfaces (APIs) and optimizations, MT6D can become a viable defense for resource constrained systems. The Python version of MT6D is analyzed initially to determine what functions might be performance bottlenecks that could be performed more efficiently using C. Based on this analysis, specific parts of the Python version are identified for improvement in the C version by either using functionality of the Linux kernel and network stack or by reworking the code in a more efficient way. After this analysis, the information gathered about the Python version is used to write the C version, using methods specific to a moving target defense to capture, analyze, modify, and tunnel packets. Finally, tests are designed and run to compare the performance of the Python and C versions.
- Masters Theses