Markus Lindgren

Using measurements to derive the worst-case execution time

Execution time analysis is used in the development of real time and embedded systems to derive the timing estimates required for schedulability analysis. The execution time of the analyzed program is typically obtained by combining results from program flow analysis (such as number of iterations in loops) with low level timing information. The paper proposes a method for low level timing analysis based on measurements of execution times of programs executing on the actual target architecture. The essence of the method is to derive a system of linear equations from a limited number of timing measurements of an instrumented version of the considered program. The solution to these equations give execution times for program fragments, from which execution time measures for the entire program can be derived. The main advantage with this approach is that architectural modeling is not needed, hence the risk of a discrepancy between model and real system is avoided. Also, compared to the non-exhaustive measurements performed in industry today, our approach is more structured and gives complete coverage in terms of the program paths considered. We present our method in the context of a simple, but reasonably realistic processor model and show how it can be extended to architectures with pipelines.

Virtualization technologies in embedded real-time systems

Virtualization is a promising solution to develop complex embedded systems with real-time requirements. This paper discusses the current state-of-the-art in virtualization technologies, with a particular focus on solutions for embedded real-time systems. Several such solutions have been developed over the past decade, and in this paper we give an overview of the more well known technologies and we provide a comparative assessment of key virtualization techniques available in these solutions. Gaps and lacking pieces are identified and further development and research is suggested.

Importance of Software Architecture during Release Planning

Release planning is the process of deciding what to include in future release(s) of a product. In this paper we look at how software architects are involved during release planning in industry today, and how architectural issues are considered during this phase.

Key Aspects of Software Release Planning in Industry

Software release planning is the process of deciding what to include in future release(s) of a product. Basically the problem can be seen as a company-wide optimization problem involving many stakeholders where the goal is to maximize utilization of the often limited resources of a company and turn them into business benefit. Saliu and Ruhe have proposed a set of key aspects for release planning methods, of which only a subset have been validated in industry. In this paper we use the Saliu and Ruhe key aspects as a starting point for identifying key aspects of release planning. To do this we have performed a multiple case study involving 7 international industrial companies, all producers of software intensive products. Our contribution is (1) a more strict meaning of a release planning key aspect, (2) validation of some of the aspects proposed by Saliu and Ruhe, and (3) an extension of the key aspects. We also capture state-of-the-practice for release planning in industry.

Towards a Capability Model for the Software Release Planning Process -- Based on a Multiple Industrial Case Study

Software release planning is an important activity for effectively identifying the customer needs generating best business, especially for incremental software development. In this paper we propose a capability model for improving the release planning process of an organization. Using this model it is possible to 1) determine the capabilities of an organizations release planning process, and 2) identify areas for improvement. The model is based on empirical data from a multiple case study involving 7 industrial companies, all being producers of software intensive systems. We also provide examples of how the proposed capability model can be applied using the companies from the study.

A Method for Balancing Short- and Long-Term Investments: Quality vs. Features

There are a number of conflicting forces between short- and long-term considerations for software release planning in industry. For example, from a business perspective it is usually desired with a short time-to-market. However, from software quality perspective it is usually desired to have a longer time-to-market such that the proper architectural mechanisms can be put in place, which in the long-term reduce development cost and addresses quality aspects. In this paper we outline some of these conflicting forces, with a focus on long-lived systems, and examplify how they impact product quality and time-to-market. In this paper we propose a simple, but useful, extension of the release planning process that addresses these conflicting forces. The method is inspired from empirical data captured in a multiple case study involving 7 companies.

GPGPU for industrial control systems

In this work in progress paper we present parts of our ongoing work on using the Graphical Processing Unit (GPU) in the context of Embedded Systems. As a first step we are investigating the possibility to move functions from a Digital Signal Processor (DSP) to a GPU. If it is possible to make such a migration then it would simplify the hardware designs for some embedded systems by removing external hardware and also remove a potential life cycle issue with obsolete components. We are currently designing a test system to be able to compare performance between a legacy control system used today in industry, based on a CPU/DSP combination, to a new design with a CPU/GPU combination. In this setting the pre-filtering of sampled data, previously done in the DSP, is moved to the GPU.

Deriving reliability estimates of distributed real-time systems by simulation

Industrial deployment of academic real-time techniques still struggles to gain momentum due to the non-familiarity of industry with schedulability analysis, as well as the lack of appropriate commercial tools. Moreover, it is imperative that academia realises the extent of the pessimism for the proposed techniques, which often makes them less attractive to systems developers. The possible trade-offs in timing guarantees vs. reliability is one such key area which needs closer study and scrutiny. There is a need for less stringent guarantees in order to avoid costly overdesigns of systems. We present a framework and simulation based methodology for reliability analysis of distributed real-time systems. We have developed a tool which is quite versatile and can accommodate varied task models, network topologies and scheduling paradigms. The tool is illustrated by a comprehensive case-study. Since our method is based on simulation, which is a standard practice in many industrial projects, we believe it will be more comprehensible and acceptable to industry.

Applicability of using internal GPGPUs in industrial control systems

Industrial control systems are continuously increasing in functionality, connectivity, and levels of integration, and as a consequence they require more computational power. At the same time, these systems have specific requirements related to cost, reliability, timeliness, and thermal power dissipation, which put restrictions on the hardware and software used. Today the high-end embedded CPUs not only provide multiple cores, but also integrated graphics processors (GPU) at close to no additional cost. The use of GPUs for general processing have several potential values in industrial control systems; 1) the added computational power and the high parallelism could pave way for new functionality and 2) the integrated GPU could potentially replace other hardware and thereby reduce the overall cost. In this paper we investigate the applicability of using integrated GPUs in industrial control systems. We do this by evaluating the performance of GPUs with respect to computational problem types and sizes typically found in industrial control systems. In the end we conclude that GPUs are no obvious match for industrial control systems and that several hurdles remain before a wide adoption can be motivated.