1 May 2012 in Communications of the ACM
Deterministic execution offers many benefits for debugging, fault tolerance, and security. Current methods of executing parallel programs deterministically, however, often incur high costs, allow misbehaved software to defeat repeatability, and transform time-dependent races into input-or path-dependent races without eliminating them. We introduce a new parallel programming model addressing these issues, and use Determinator, a proof-of-concept OS, to demonstrate the model's practicality. Determinator's microkernel application programming interface (API) provides only "shared-nothing" address spaces and deterministic interprocess communication primitives to make execution of all unprivileged code---well-behaved or not---precisely repeatable. Atop this microkernel, Determinator's user-level runtime offers a private workspace model for both thread-level and process-level parallel programming. This model avoids the introduction of read/write data races, and converts write/write races into reliably detected conflicts. Coarse-grained parallel benchmarks perform and scale comparably to non-deterministic systems, both on multicore PCs and across nodes in a distributed cluster.
1 January 2012 in Lecture Notes in Computer Science
The consequences of security breaches due to system administrator errors can be catastrophic. Software systems in general, and OSes in particular, ultimately depend on a fully trusted administrator whom is granted superuser privileges that allow him to fully control the system. Consequently, an administrator acting negligently or unethically can easily compromise user data in irreversible ways by leaking, modifying, or deleting data. In this paper we propose a new set of guiding principles for OS design that we call the broker security model. Our model aims to increase OS security without hindering manageability. This is achieved by a two-step process that (1) restricts administrator privileges to preclude inspection and modification of user data, and (2) allows for management tasks that are mediated by a layer of trusted programs—brokers—interposed between the management interface and system objects. We demonstrate the viability of this approach by building BrokULOS, a Linux-based OS that suppresses superuser privileges and exposes a narrow management interface consisting of a set of tailor-made brokers. Our evaluation shows that our modifications to Linux add negligible overhead to applications while preserving system manageability.