PhD Proposal: Execution Environments for Running Legacy Applications in Multi-Party Trust Settings

Applications often assume a monolithic trust setting where a single party controls the ap- plication, its data, and its execution environment. This assumption no longer holds when organizations outsource applications to a third-party cloud, or use the application to compute over a heterogeneous dataset, such as data from multiple parties. The result of this broken assumption is that using applications in these settings leaks one party’s private data to another. I propose to evaluate the following thesis: it is possible to run legacy application binaries with confidentiality and integrity guarantees that reflect a multi-party trust setting. My approach is to apply operating system designs and fine-grained information flow control to hoist, or otherwise partition, the execution environment into trust boundaries in a manner that is transparent to the application.In the first part of my proposal, I review my prior work in extending a library operating system that runs within an Intel SGX secure hardware enclave, so as to support running a broader set of trusted, legacy, applications in untrusted environments. In the second part, I discuss my proposed work, codomains, an execution model that maintains the source-level abstraction of a monolithic program, but allows an application to dynamically switch execution to different domains (hosts and enclaves) via language-neutral mechanisms.Examining Committee:

Chair: Dr. Dave Levin Dept rep: Dr. David Van Horn Members: Dr. Bobby Bhattacharjee Dr. Christina Garman Dr. Deepak Garg