Skip to Main Content
Distributed hard real-time systems require predictable communication at the network level and verifiable communication behavior at the application level. At the network level, communication between nodes must be guaranteed to happen within bounded time and one common approach is to restrict the network access by enforcing a time-division multiple access (TDMA) schedule. At the application level, the application's communication behavior should be verified to ensure that the application uses the predictable communication in the intended way. Network code is a domain-specific programming language to write a predictable verifiable distributed communication for distributed real-time applications. In this paper, we present the syntax and semantics of network code, how we can implement different scheduling policies, and how we can use tools such as model checking to formally verify the properties of network code programs. We also present an implementation of a runtime system for executing network code on top of RTLinux and measure the overhead incurred from the runtime system.