Real-Time Performance and Middleware for Multiprocessor and Multicore Linux Platforms

An increasing number of distributed real-time applications are running on multicore platforms. However, existing real-time middleware (e.g., Real-Time CORBA) lacks adequate support for ensuring the timing constraints of soft real-time tasks on multicore platforms, and thus is dependent on (potentially inadequate) support from the underlying operating system. This paper makes three contributions to the state of the art in real-time system software for multicore platforms. First, it offers what is to our knowledge the first experimental analysis of real-time performance of vanilla Linux primitives on multicore platforms. Second, it presents MC-ORB, the first real-time object request broker (ORB) designed to address the nuances of multiprocessor (and especially multicore) platforms with a novel core-aware middleware thread architecture and allocation service for soft real-time tasks. Third, it evaluates MC-ORB's performance on a Linux multicore testbed, the results of which demonstrate its efficiency and effectiveness.