Mikael Nolin

SIRAP: a synchronization protocol for hierarchical resource sharingin real-time open systems

This paper presents a protocol for resource sharing in a hierarchical real-time scheduling framework. Targeting real-time open systems, the protocol and the scheduling framework significantly reduce the efforts and errors associated with integrating multiple semi-independent subsystems on a single processor. Thus, our proposed techniques facilitate modern software development processes, where subsystems are developed by independent teams (or subcontractors) and at a later stage integrated into a single product. Using our solution, a subsystem need not know, and is not dependent on, the timing behaviour of other subsystems; even though they share mutually exclusive resources. In this paper we also prove the correctness of our approach and evaluate its efficiency.

The Rubus component model for resource constrained real-time systems

In this paper we present a component model for development of distributed real-time systems. The model is developed to support development of embedded control systems for ground vehicles. The model aims at supporting three important activities in real-time development, (i) design, (ii) analysis and (iii) synthesis. These activities emphasise different and sometimes conflicting requirements that need to be balanced. For example, developers desire freedom in designing to solve complex tasks, analysis tools require the design to be formal enough for analysis and synthesis need to be efficient for low run-time footprint. We have considered industrial requirements for these activities and developed the RubusCMv3 component model. The model has been developed in close cooperation with industrial partners and it is currently being evaluated on real systems.

Real-time server-based communication with CAN

This paper investigates the concept of share-driven scheduling of networks using servers with real-time properties. Share-driven scheduling provides fairness and bandwidth isolation between predictable as well as unpredictable streams of messages on the network. The need for this kind of scheduled real-time communication network is high in applications that have requirements on flexibility, both during development for assigning communication bandwidth to different applications, and during run-time to facilitate dynamic addition and removal of system components. We illustrate the share-driven scheduling concept by applying it to the popular controller area network (CAN). We propose a scheduling mechanism that we call simple server-scheduled CAN (S/sup 3/-CAN), for which we also present an associated timing analysis. Additionally, we present a variant of S/sup 3/-CAN called periodic server-scheduled CAN (PS/sup 2/-CAN), which for some network configurations gives lower worst-case response-times than S/sup 3/-CAN. Also for this improvement, a timing analysis is presented. Moreover, we use simulation to evaluate the timing performance of both S/sup 3/-CAN and PS/sup 2/-CAN, comparing them with other scheduling mechanisms.

Asymmetric multihop communication in large sensor networks

With the growing interest in wireless sensor networks, energy efficient communication infrastructures for such networks are becoming increasingly important. In this paper, we compare and simulate asymmetric and symmetric communication in sensor networks. We do this by extending LEACH, a well-known TDMA cluster-based sensor network architecture, to use asymmetric communication. The extension makes it possible to scale up the network size beyond what is feasible with LEACH and its variants LEACH-C and LEACH-F.

Synthesis of Optimal Interfaces for Hierarchical Scheduling with Resources

This paper presents algorithms that (1) facilitate system-independent synthesis of timing-interfaces for subsystems and (2) system-level selection of interfaces to minimize CPU load. The results presented are developed for hierarchical fixed-priority scheduling of subsystems that may share logical recourses (i.e. semaphores). We show that the use of shared resources results in a tradeoff problem, where resource locking times can be traded for CPU allocation, complicating the problem of finding the optimal interface configuration subject to schedulability. This paper presents a methodology where such a tradeoff can be effectively explored. It first synthesizes a bounded set of interface-candidates for each subsystem, independently of the final system, such that the set contains the interface that minimizes system load for any given system. Then, integrating subsystems into a system, it finds the optimal selection of interfaces. Our algorithms have linear complexity to the number of tasks involved. Thus, our approach is also suitable for adaptable and reconfigurable systems.

Evaluation of component technologies with respect to industrial requirements

We compare existing component technologies for embedded systems with respect to industrial requirements. The requirements are collected from the vehicular industry, but our findings are applicable to similar industries developing resource constrained safety critical embedded distributed real-time computer systems. One of our conclusions is that none of the studied technologies is a perfect match for the industrial requirements. Furthermore, no single technology stands out as being a significantly better choice than the others; each technology has its own pros and cons. The results of our evaluation can be used to guide modifications or extensions to existing technologies, making them better suited for industrial deployment. Companies that want to make use of component-based software engineering as available today can use this evaluation to select a suitable technology.

Industrial Requirements on Component Technologies for Embedded Systems

Software component technologies have not yet been generally accepted by embedded-systems industries. In order to better understand why this is the case, we present a set of requirements, based on industrial needs, that are deemed decisive for introducing a component technology. The requirements we present can be used to evaluate existing component technologies before introducing them in an industrial context. They can also be used to guide modifications and/or extensions to component technologies, to make them better suited for industrial deployment. One of our findings is that a major source of requirements is non-technical in its nature. For a component technology to become a viable solution in an industrial context, its impact on the overall development process needs to be addressed. This includes issues like component life-cycle management, and support for the ability to gradually migrate into the new technology.

Contract-Based ReusableWorst-Case Execution Time Estimate

We present a contract-based technique to achieve reuse of known worst-case execution times (WCET) in conjunction with reuse of software components. For resource constrained systems, or systems where high degree of predictability is needed, classical techniques for WCET- estimation will result in unacceptable overestimation of the execution-time of reusable software components with rich behavior. Our technique allows different WCETs to be associated with subsets of the component behavior. The appropriate WCET for any usage context of the component is selected be means of component contracts over the input domain. In a case-study we illustrate our technique and demonstrate its potential in achieving tight WCET- estimates for reusable components with rich behavior.

Pessimistic concurrency control and versioning to support database pointers in real-time databases

In this paper we present a concurrency control algorithm that allows co-existence of soft real-time, relational database transactions, and hard real-time database pointer transactions in real-time database management systems. The algorithm uses traditional pessimistic concurrency-control (i.e. locking) for soft transactions and versioning for hard transactions to allow them to execute regardless of any database lock. We provide formal proof that the algorithm is deadlock free and formally verify that transactions have atomic semantics. We also present an evaluation that demonstrates significant benefits for both soft and hard transactions when our algorithm is used. The proposed algorithm is suited for resource-constrained safety critical, real-time systems that have a mix of hard real-time control applications and soft real-time management, maintenance, or user-interface applications.

Fast and tight response-times for tasks with offsets

In previous work, we presented a tight approximate response-time analysis for tasks with offsets. While providing a tight bound on response times, the tight analysis exhibits similarly long execution times as does the traditional methods for calculating response-times for tasks with offsets. The existing method for fast analysis of tasks with offsets is not applicable to the tight analysis. In this paper we extend the fast analysis to handle the distinguishing trait of the tight analysis; continuously increasing interference functions. Furthermore, we provide another speedup; by introducing pessimism in the modelling of interference at certain points, we speed up the convergence of the numerical solving for response-times without increasing the pessimism of the resulting response-times. The presented fast-and-tight analysis is guaranteed to calculate the same response-times as the tight analysis, and in a simulation study we obtain speedups of more than two orders of magnitude for realistically sized tasks sets compared to the tight analysis. We also demonstrate that the fast-and-tight analysis has comparable execution time to that of the fast analysis. Hence, we conclude that the fast-and-tight analysis is the preferred analysis technique when tight estimates of response-times are needed, and that we do not need to sacrifice tightness for analysis speed; both are obtained with the fast-and-tight analysis.