Evaluation of popular application sandboxing

Sandboxing is a modern technique that can be employed at the application or OS level to test untrusted or untested programs. It monitors and limits the level of access that the program has. This can prevent the damaging effect of a program if it is malicious or not coded properly. Nonetheless, sandboxing adds extra overhead cost which affects the speed and performance of a running program. In this paper, we focus on evaluating experimentally popular sandboxes that run at the application level. We examine the performance of application sandboxes by running different types of benchmarks within these sandboxes. We examine the performance in terms of execution time and the usage of physical memory, disk I/O, and network I/O.