Johan Kraft

A Metaheuristic Approach for Best Effort Timing Analysis Targeting Complex Legacy Real-Time Systems

Many companies developing real-time systems today have today no means for response time analysis, as their systems violate the assumptions of traditional analytical methods for response-time analysis and are too complex for exhaustive analysis using model checking. This paper presents a novel approach for best effort response time analysis targeting such systems, where probabilistic simulation is guided by a search algorithm of metaheuristic type, similar to genetic algorithms. The best effort approach means that the result is not guaranteed to be the worst-case response time, but also that the method scales to large industrial systems. The proposed method should be regarded as a form of testing, focusing on timing properties. An evaluation is presented which indicates that the proposed approach is significantly more efficient than traditional probabilistic simulation in finding extreme task response times. The paper also presents a method for finding good parameters for the search algorithm, in order to improve its efficiency.

Trace recording for embedded systems: lessons learned from five industrial projects

This paper presents experiences from five industry collaboration projects performed between 2004-2009 where solutions for embedded systems trace recording have been developed and evaluated; in four cases for specific industrial systems and in the last case as a generic solution for a commercial real-time operating system, in collaboration with the RTOS company. The experiences includes technical solutions regarding efficient instrumentation and logging, technology transfer issues and evaluation results regarding CPU and RAM overhead. A brief overview of the Tracealyzer tool is also presented, a result of the first project (2004) which still is used by ABB Robotics and now in commercialization.

Simulation-Based Timing Analysis of Complex Real-Time Systems

This paper presents an efficient best-effort approach for simulation-based timing analysis of complex real-time systems. The method can handle in principle any software design that can be simulated, and is based on controlling simulation input using a simple yet novel hill-climbing algorithm. Unlike previous approaches, the new algorithm directly manipulates simulation parameters such as execution times, arrival jitter and input. An evaluation is presented using six different simulation models, and two other simulation methods as reference: Monte Carlo simulation and MABERA. The new method proposed in this paper was 4-11 % more accurate while at the same time 42 times faster, on average, than the reference methods.

A Statistical Approach to Response-Time Analysis of Complex Embedded Real-Time Systems

This paper presents Rapid RT, a novel statistical approach to Worst-Case Response-Time (WCRT) analysis targeting complex embedded real-time systems. The proposed algorithm combines Extreme Value Theory (EVT) and other statistical methods in order to produce a probabilistic WCRT estimate. This estimate is calculated using response time data from either Monte Carlo simulations of a detailed model of the system, or from response-time measurements of the real system. The method could be considered as a pragmatic approach intended for complex industrial systems with real-time requirements. The target systems contain tasks with many intricate dependencies in their temporal behavior, which violates the assumptions of traditional analytical methods for response time analysis and thereby makes them overly pessimistic. An evaluation is presented using two simulation models, inspired by an industrial robotic control system, and five other methods as reference.

Towards migrating legacy real-time systems to multi-core platforms

Power consumption and thermal problems limit the single-core processors to be faster. Processor architects are therefore moving toward multi-core processors. Developers of embedded real-time systems however hesitates a shift to multi-core processors, especially for existing ldquolegacyrdquo systems which have been developed with single-core processor assumptions. These systems have been developed and maintained by many developers over many years, and can not easily be replaced due to the huge development investments they represent. In this paper we investigate challenges of migrating complex legacy real-time systems to multi-core architectures. We propose componentization and partitioning to prepare the migration. Componentization groups logically related tasks into components (or subsystems). This provides an abstraction layer from a scheduling perspective, which facilitates migration. Partitioning maps tasks to the different cores on the multi-core processor, maximizing system performance while ensuring correctness.

A statistical approach to simulation model validation in response-time analysis of complex real-time embedded systems

As simulation-based analysis methods make few restrictions on the system design and scale to very large and complex systems, they are widely used in, e.g., timing analysis of complex real-time embedded systems (CRTES) in industrial circles. However, before such methods are used, the analysis simulation models have to be validated in order to assess if they represent the actual system or not, which also matters to the confidence in the simulation results. This paper presents a statistical approach to validation of temporal simulation models extracted from CRTES, by introducing existing mature statistical hypothesis tests to the context. Moreover, our evaluation using simulation models depicting a fictive but representative industrial robotic control system indicates that the proposed method can successfully identify temporal differences between different simulation models, hence it has the potential to be considered as an effective simulation model validation technique.

Statistical-Based Response-Time Analysis of Systems with Execution Dependencies between Tasks

This paper presents a novel statistical-based approach to Worst-Case Response-Time (WCRT) analysis of complex real-time system models. These system models have been tailored to capture intricate execution dependencies between tasks, inspired by real industrial control systems. The proposed WCRT estimation algorithm is based on Extreme Value Theory (EVT) and produces both WCRT estimates together with a probability of being exceeded. By using the tools developed, an evaluation is presented using three different simulation models, and four other methods as reference: Monte Carlo simulation, MABERA, HCRR and traditional Response-Time Analysis (basic RTA). Empirical results demonstrate that the benefit of the proposed approach, in terms of 1) reduced pessimism when compared to basic RTA and 2) validated guarantee of never being less than the actual response time values. The proposed approach also needs much fewer simulations compared to other three simulation-based methods.

Software Reverse Engineering in the Domain of Complex Embedded Systems

This chapter provides a review of reverse engineering of software for complex embedded systems. Our review is motivated by the observation that the reach and importance of embedded systems are grow ...

Transformational Specification of Complex Legacy Real-Time Systems via Semantic Anchoring

RTSSim is a framework for simulating models extracted from complex legacy real-time systems which are task-oriented, run on a single processor and are developed in C. Such RTSSim models describe functional and temporal behavior as well as the resource usage of the system. However, the semantics specification of RTSSim models remains a challenging problem indeed, especially with tractable complexity to obtain a formal model which can be analyzed for instance by a model checking tool. In this paper, we present an approach towards using semantic anchoring for the transformational specification of RTSSim models, by relying on units with well-defined operational semantics and tool support. Specifically, Timed Automata with Tasks (TAT) in TIMES is chosen as the semantic unit with the purpose of anchoring different behavioral concerns of RTSSim models in all aspects. In this respect, model transformations are conducted at the meta-model level allowing the original operational semantics of RTSSim models to be preserved, while at the same time it can be presented in TIMES models in terms of a network of TAT.

Evaluating the Quality of Models Extracted from Embedded Real-Time Software

Due to the high cost of modeling, model-based techniques are yet to make their impact in the embedded systems industry, which still persist on maintaining code-oriented legacy systems. Re-engineering existing code-oriented systems to fit model-based development is a risky endeavor due to the cost and efforts required to maintain correspondence between the code and model. We aim to reduce the cost of modeling and model maintenance by automating the process, thus facilitating model-based techniques. We have previously proposed the use of automatic model extraction from recordings of existing embedded real-time systems. To estimate the quality of the extracted models of timing behavior, we need a framework for objective evaluation. In this paper, we present such a framework to empirically test and compare extracted models, and hence obtain an implicit evaluation of methods for automatic model extraction. We present a set of synthetic benchmarks to be used as test cases for emulating timing behaviors of diverse systems with varying architectural styles, and extract automatic models out of them. We discuss the difficulties in comparing response time distributions, and present an intuitive and novel approach along with associated algorithms for performing such a comparison. Using our empirical framework, and the comparison algorithms, one could objectively determine the correspondence between the model and the system being modeled