A Novel Framework For Executing Untrusted Programs

Lot of studies has been done on tackling untrusted programs. Most of the studies are aimed at preventing an untrusted program from execution. Models have been established to provide proofs to codes so that a program is allowed to execute only when the program passes a test. The test may be in the form of some proof, certificate etc. The basic idea is not to allow an untrusted program to execute if the safety cannot be guaranteed. This paper proposes a totally novel approach to tackle untrusted programs, infact it is diametrically opposite to the existing policies. It starts with the assumption that all untrusted programs are safe until proven otherwise. The best way to prove the safety of a program is to execute it at least once. We have proposed a new model to execute such programs in a virtual environment and then decide on the trustworthiness of the program.