1 Abstract 1 The increasing acceptance of the Earliest Deadline First (EDF) scheduling algorithm in industrial environments, together with the continued usage of Fixed Priority (FP) scheduling is leading to heterogeneous systems with different scheduling policies in the same distributed system. Schedulability analysis techniques usually consider the entire system as a whole (holistic approach), with only one preestablished scheduling policy in all the resources. In this work, composition mechanisms will be proposed that enable us to combine different FP and EDF response-time analysis techniques for checking the schedulability of heterogeneous systems. Additionally, priority and scheduling deadline assignment techniques will be combined into the a new algorithm called HOSPA (Heuristic Optimized Scheduling Parameters Assignment), for optimizing the assignment of priorities and scheduling deadlines to tasks and messages in heterogeneous distributed hard real-time systems. 1. Introduction 2 The usage of Earliest Deadline First (EDF) scheduling policy is starting to catch the attention of industrial environments, given its benefits in terms of increased resource usage. EDF is now present at different layers of a real-time application such as programming languages, operating systems, or even communication networks. So, it is available in real-time languages like Ada 2005 [26] or RTSJ [20], and in real-time operating systems such as SHaRK [21] or ERIKA [3]. It has been also implemented at the application level in OSEK/VDX embedded operating systems [1], and there are real-time networks using EDF for scheduling messages too; for instance in general purpose networks [2], or in the CAN Bus [17]. It is expected that the number of industrial applications using EDF will increase in the foreseeable future. On the other hand, Fixed Priority (FP) scheduling continues to be the most popular on-line scheduling policy. Depending on the particular requirements, for instance a mixture of hard real-time requirements together with requirements for different criticality levels, one or another scheduling policy may lead to better results. In a distributed system the scheduling problem can be better solved if different scheduling algorithms could be combined, so each processing resource would be scheduled with the optimum algorithm for the work it’s going to handle. There may also be other reasons to pick a particular scheduler for certain parts of the system, for instance we may have some legacy parts designed for FP, or an FP network such as a CAN bus, integrated with other processors where EDF is chosen because of its better usage of the available resources. Existing response-time analysis and scheduling parameters assignment techniques that operate with the whole system are defined only for homogeneous systems, where every processor and network is scheduled with FP or EDF, so these techniques are not used for heterogeneous systems where FP and EDF schedulers could coexist. Different abstractions have been proposed to make the performance analysis of distributed hard real-time systems. A comparison among some of these performance analysis techniques for distributed systems has been presented in [18]. That work distinguishes four different abstractions for formal performance analysis of distributed embedded systems in early design stages: holistic scheduling (MAST, Modeling and Analysis Suite for Real-Time Applications[4][12]), two compositional analysis approaches (SymTA/S [6] and MPA-RTC [24]), and a timed automata based analysis. The compositional approaches combine local analysis and output event propagation models allowing the formal performance analysis of heterogeneous distributed systems which use different scheduling algorithms in their processors and communication networks, supporting the 1. This paper is a draft version of the one submitted to the 23rd Euromicro Conference on Real-Time Systems to be held in Porto, Portugal, 6th-8th of July of 2011. 2. This work has been funded in part by the Spanish Ministry of Science and Technology under grant number TIN2008-06766-C03-03 (RT-MODEL), and by the European Commission under project FP7/NoE/ 214373 (ArtistDesign). Combining EDF and FP in Distributed Real-Time Systems: Schedulability Analysis and Optimization Juan M. Rivas, J. Javier Gutiérrez, J. Carlos Palencia, and Michael González Harbour Computers and Real-Time Group, Universidad de Cantabria, 39005-Santander, SPAIN {rivasjm, gutierjj, palencij, mgh}@unican.es