Adaptation, Power Consumption and Scheduling

Energy Management for Embedded Multithreaded Processors with Integrated EDF Scheduling

Abstract

This paper proposes a new hardware-based energy management technique for future embedded multithreaded processors with integrated Earliest Deadline First (EDF) real-time scheduling. Our energy management technique controls frequency reduction and dynamic voltage scaling depending on the deadlines, the Worst Case Execution Times (WCET), and the real execution times. Hard real-time capability can be guaranteed for aperiodic threads and for threads with deadlines shorter than their period. Our evaluations show that energy consumption can be reduced up to about \(\frac{2}{3}\) of a comparable software-based algorithm.

Sascha Uhrig, Theo Ungerer

Reducing System Level Power Consumption for Mobile and Embedded Platforms

Abstract

The power consumption of peripheral devices is a significant portion of the overall energy usage of a mobile platform. To take advantage of idle times, most devices offer the ability to transition into low power states. However, the amount of energy saved by utilizing these sleep states depends on the lengths and number of idle periods experienced by the device. This paper describes a new process scheduling algorithm which accumulates device usage information in the form of device windows to make power a first class resource: it attempts to increase the burstiness of both device accesses and idle periods, and it provides enhanced behavior for timeout-based sleep mechanisms. An initial implementation based on the default Linux scheduler demonstrates the algorithm’s and approach’s ability to reduce the average power consumption of devices by increasing device sleep times and reducing transition overheads.

Ripal Nathuji, Karsten Schwan

Implementing Control Algorithms Within a Multithreaded Java Microcontroller

Abstract

Our aim is to investigate if it is possible to control the throughput (IPC rate) of a thread running on a multithreaded Java processor by a closed feedback loop. We implemented a Proportional/Integral/Differential (PID) controller in the processor simulator of the Komodo microcontroller developed at the universities of Karlsruhe and Augsburg to simulate the PID controller as an additional hardware module. It uses the GP (Guaranteed Percentage) scheduling to control the thread. Evaluations show that the aimed IPC rate of a thread is achieved by the controller thus improving the real-time capabilities of the Java processor.

Uwe Brinkschulte, Mathias Pacher

Adaptivity for Quality and Timeliness Flexible Real-Time Systems

Abstract

The basis for this work is a model for fine-granular flexibility of applications in two directions. These are the quality of computations on the one hand and their timeliness on the other hand. Dynamic scheduling of quality- and timeliness-flexible tasks on the same hardware platform as the application itself exhibits two obvious sources of trade-offs. The first one exists between the desired quality levels for individual tasks (depending on the processing time awarded to them) and the ability of these tasks to meet timing constraints. The second one can be found between the overall distribution of processing time between the application tasks and the scheduling algorithm. A high processing time allowance granted to the scheduler may leave too little resources for the actual application; however, a small scheduling allowance might prevent finding good schedules according to the given objective function. We use a control-theoretic approach to allow the scheduler to adapt to the current characteristics of the application in terms of workload and frequency and regularity of task releases automatically at run-time.

Thomas Schwarzfischer

Adaptation and Agents

Apricot Agent Platform for User-Friendly Mobile Service Development

Abstract

The trend of increasing mobile services has set a new challenge for service providers: how to develop more advanced and user friendly, context-aware and personalized mobile services for the users. Apricot Agent Platform is an agent-based software platform designed for developing context-aware and personalized mobile services. Apricot Agent Platform supports the development of user-friendly mobile services by providing tools for combining various mobile and Internet-based services. Apricot agent architecture consists of an agent platform, agents and agent containers. For the developers of mobile services, it provides built-in functionality and communication mechanism. Furthermore, this paper describes four demonstrators that are built up on the Apricot Agent Platform to evaluate the usability and efficiency of the platform in processes of building mobile services. The results of the evaluation indicate relatively promising results and the further target for development is revealed.

Petteri Alahuhta, Henri Löthman, Heli Helaakoski, Arto Koskela, Juha Röning

Support of Reflective Mobile Agents in a Smart Office Environment

Abstract

Ubiquitous systems will integrate computers invisibly and unobtrusively in everyday objects. Information will appear in new forms, i.e. data will be catched from single or multi-sensor devices and will be used for context extraction. New location-based services will be adapted to user preferences. For this the ubiquitous system needs to know user profiles, likings, and habits. As the user moves these information have to be sent to the new location of the user. Either the user carries her data on wearable or portable computers or the ubiquitous environment takes responsibility for transporting them. The amount of new devices and services makes an efficient use by centralized systems very difficult.

The idea presented in this paper is that a virtual reflection of the user represented by a mobile agent accompanies her in the ubiquitous environment. Mobile agents offer a possibility to encapsulate information of a person and her preferences and perform location-based services of the ubiquitous system in the name of the user. Because of the personal data security and privacy are major concerns of such an agent system. This paper describes a ubiquitous mobile agent system named UbiMAS which has security extensions to provide high protection of agents and significant personal data. UbiMAS is used in a smart office environment with smart doorplates.

Faruk Bagci, Holger Schick, Jan Petzold, Wolfgang Trumler, Theo Ungerer

Learning Action Sequences Through Imitation in Behavior Based Architectures

Abstract

In this paper a new architecture for learning action sequences through imitation is proposed. Imitation occurs by means of observing and applying sequences of basic behaviors. When an agent has observed another agent and applied the observed action sequence later on, this imitated action sequence can be seen as a meme. Agents that behave similarly can therefore be grouped by their typical behavioral patterns. This paper thus explores imitation from the view of memetic proliferation.

Combining imitation learning with meme theory we show by simulating agent societies that with imitation significant performance improvements can be achieved. The performance is quantified by using an entropy measure to qualitatively evaluating the emerging clusters.

Our approach is demonstrated by the example of a society of emotion driven agents that imitate each other to reach pleasant emotional state.

Willi Richert, Bernd Kleinjohann, Lisa Kleinjohann

Adaptation and Services

Self-healing Execution of Business Processes Based on a Peer-to-Peer Service Architecture

Abstract

The automated execution of business processes that are composed of individual web services has seen a growing importance throughout enterprise computing in the recent years. The Business Process Execution Language for Web Services (BPEL4WS) has become the predominant language to express such business process compositions. In this paper we present the design and implementation of a Robust Execution Layer that acts as a transparent, configurable add-on to any BPEL4WS execution engine to support self-healing execution of business processes. Resilience of the process execution is achieved through service replacement in case of communication failures, by relying on a robust peer-to-peer service discovery and selection mechanism for alternative services.

Thomas Friese, Jörg P. Müller, Bernd Freisleben

Runtime Adaptation of Applications Through Dynamic Recomposition of Components

Abstract

Software applications executing in highly dynamic environments are faced with the challenge of frequent and usually unpredictable changes in their execution environment. In order to cope with this challenge effectively, the applications need to adapt to these changes dynamically. CASA (Contract-based Adaptive Software Architecture) provides a framework for enabling dynamic adaptation of applications, in response to changes in their execution environment. One of the principle adaptation mechanisms employed in the CASA framework is dynamic recomposition of application components. In this paper, we discuss implementation issues related to the approach for dynamic recomposition of application components in CASA.

Arun Mukhija, Martin Glinz

An Observer/Controller Architecture for Adaptive Reconfigurable Stacks

Abstract

In this paper, we discuss the necessity of new observation and control structures for organic computing systems starting from the basic contradiction between bottom-up behaviour and top-down design. An Observer/Controller architecture serves the purpose to keep emergent behaviour within predefined limits. As an illustration, a framework for reconfigurable protocol stacks is introduced, which contains an agent-based monitoring framework as well as a reconfiguration manager. After describing a TCP/IP protocol stack implementation, based on the framework, similarities between the introduced framework and the Observer/Controller architectural pattern will be pointed out.

Thorsten Schöler, Christian Müller-Schloer

Application Adaptable Systems

The Organic Features of the AMIDAR Class of Processors

Abstract

In this contribution we present a novel general model for adaptive processors with organic features. We describe its basic principle of operation. The adaptive operations that are possible with this model are thoroughly discussed with respect to organic computing. The model allows runtime variations of the type and number of functional units as well as variations of the communication structure. Experimental results show that a processor implementing this model can self-optimize its architecture for several diverse applications.

Stephan Gatzka, Christian Hochberger

Reusable Design of Inter-chip Communication Interfaces for Next Generation of Adaptive Computing Systems

Abstract

The SoC (System-on-Chip) technology is used in small and flexible consumer electronic devices. SoCs include one or more microcontroller, memory, programmable logic, and the input/output logic control. Additionally, sophisticated SoCs support partial dynamic reconfiguration. Those are preconditions to build the next generation of adaptive computing systems which make it possible to implement selforganizing systems that are self-configuring and self-optimizing. The design of applications and the development of tools for system design are a great challenge. In this paper we describe an approach that is used to support the design of applications by generator tools. This approach allows the re-use and the generation of communication interfaces between the components in partial run-time reconfiguration (pRTR) systems. The generator tool approach based on a methodology which enables a formal representation of adaptive systems and its timing schedule control. We prove our methodology and generator approach by applications from the field of signal processing.

Vincent Kotzsch, Jörg Schneider, Günther Döring

DESCOMP: A New Design Space Exploration Approach

Abstract

In this paper, we introduce a new approach in Design-Space-Exploration (DSE) for non-clustered VLIW architectures. It differs from existing techniques by using a “bottom-up” strategy. While other approaches start with the design of an architecture, followed by building a possible schedule, we firstly build a schedule and after that an architecture is synthesized, which is suitable to execute this schedule. So, the results can be obtained fully automatically and in very short time. Furthermore, we can explore arbitrary types of functional units without increasing the design space exploration time significantly. We evaluated our method and compared the obtained results to an existing DSE approach for clustered and non-clustered architectures. We almost always obtain better results in the case of non-clustered architectures. In many cases the ports of the register file are decreased, which, in consequence, leads to higher clock rates. Compared to the results for clustered architectures for some examples our non-clustered architecture is better than the best clustered one.

Mario Schölzel, Peter Bachmann

Design Space Navigation for Neighboring Power-Performance Efficient Microprocessor Configurations

Abstract

Microprocessor design is a considerably complex task. First, microprocessors include many resources that may be configured in different ways. This leads to a time consuming multi-objective optimization problem. Second, currently the designs must take into account not only performance but also power consumption thus making the optimization goal more complex. Third, different types of applications have different demands but producing several different microprocessors would not be cost effective.

This paper proposes an efficient algorithm to explore the design space: design space navigation. With this algorithm it is possible to obtain optimal configurations by starting from a baseline and “navigating” on the design space. Different configurations tailored for different applications, but derived from the same baseline, are called neighboring configurations. Experimental results show that navigation finds designs that achieve better power-performance efficiency for a fraction of the time required by other design space exploration algorithms. Also, the algorithm is used to obtain four neighboring configurations for four types of applications: multimedia, integer and floating-point scientific, and database workloads. The results showed that the navigation configuration achieves a power-performance improvement of 30% to 118% depending on the workload. Using different workloads for navigation and execution may result in a loss of efficiency of as much as 94%.

Pedro Trancoso

An Efficient Frequency Scaling Approach for Energy-Aware Embedded Real-Time Systems

Abstract

The management of energy consumption in battery-operated embedded and pervasive systems is increasingly important in order to extend battery lifetime or to increase the number of applications that can use the system’s resources. Dynamic voltage and frequency scaling (DVFS) has been introduced to trade off system performance with energy consumption. For real-time applications, systems supporting DVFS have to balance the achieved energy savings with the deadline constraints of applications. Previous work has used periodic evaluation of an application’s progress (e.g., with periodic checkpoints inserted into application code at compile time) to decide if and how much to adjust the frequency or voltage. Our approach builds on this prior work and addresses the overheads associated with these solutions by replacing periodic checkpoints with iterative checkpoint computations based on predicted best-, average-, and worst-case execution times of real-time applications (e.g., obtained through compile-time analysis or profiling).

Christian Poellabauer, Tao Zhang, Santosh Pande, Karsten Schwan

Pervasive Computing and Communication

Towards Autonomic Networking Using Overlay Routing Techniques

Abstract

With an ever-growing number of computers being embedded into our surroundings, the era of ubiquitous computing is approaching fast. However, as the number of networked devices increases, so does system complexity. Contrary to the goal of achieving an “invisible computer”, the required amount of management and human intervention increases more and more, both slowing down the growth rate and limiting the achievable size of ubiquitous systems.

In this paper we present a novel routing approach that is capable of handling complex networks without any administrative intervention. Based on a combination of standard overlay routing techniques and source routes, this approach is capable of efficiently bootstrapping a routable network. Unlike other approaches that try to combine peer-to-peer ideas with ad-hoc networks, sensor networks, or ubiquitous systems, our approach is not based on a routing scheme. This makes the resulting system flexible and powerful with respect at application support as well as efficient with regard to routing overhead and system complexity.

Kendy Kutzner, Kurt Cramer, Thomas Fuhrmann

Context-Based Storage Management for Wearable and Portable Devices

Abstract

In our information-rich world, managing the data we collect is becoming a significant bottleneck for users. This issue has triggered considerable research in so-called semantic file systems, relying on the attachment of meta-data to files. Such data is useful for dynamically arranging files in virtual directories, according to the user’s request or task at hand. However, current research typically concerns the desktop and little work has been done taking into account mobile devices, which in addition to being generators of data themselves are now also capable of carrying significant amounts of information. In this paper we discuss how context information can be exploited to attach semantics to files residing on portable devices, and we show how such information can be used by the storage system itself to enhance data management while on the move. We also describe a storage framework which integrates smoothly with semantic file systems by facilitating automatic annotation of files generated by portables, as well as by exploiting this information to better integrate with infrastructure data stores.

Alexandros Karypidis, Spyros Lalis

A File System for System Programming in Ubiquitous Computing

Abstract

In Ubiquitous computing small embedded sensor and computing nodes are one of the main enabling technologies. System programming for such small embedded systems is a challenging task involving various hardware components with different characteristics. This paper presents a file system for sensor nodes platforms providing a common organization structure and a lightweight and uniform access model for sensors and all other resources on sensor nodes. This mechanism forms an abstraction from different hardware, makes functions re-useable and simplifies the development on such systems. With ParticleFS an file system implementation on a sensor node platform is shown. As an example a telnet application running on sensor nodes was implemented demonstrating the usage of the approach for system programming on such platforms.

Christian Decker, Michael Beigl, Albert Krohn

Springer Professional

Systems Aspects in Organic and Pervasive Computing - ARCS 2005

18th International Conference on Architecture of Computing Systems, Innsbruck, Austria, March 14-17, 2005. Proceedings

Inhaltsverzeichnis

Frontmatter

Adaptation, Power Consumption and Scheduling

Energy Management for Embedded Multithreaded Processors with Integrated EDF Scheduling

Reducing System Level Power Consumption for Mobile and Embedded Platforms

Implementing Control Algorithms Within a Multithreaded Java Microcontroller

Adaptivity for Quality and Timeliness Flexible Real-Time Systems

Adaptation and Agents

Apricot Agent Platform for User-Friendly Mobile Service Development

Support of Reflective Mobile Agents in a Smart Office Environment

Learning Action Sequences Through Imitation in Behavior Based Architectures

Adaptation and Services

Self-healing Execution of Business Processes Based on a Peer-to-Peer Service Architecture

Runtime Adaptation of Applications Through Dynamic Recomposition of Components

An Observer/Controller Architecture for Adaptive Reconfigurable Stacks

Application Adaptable Systems

The Organic Features of the AMIDAR Class of Processors

Reusable Design of Inter-chip Communication Interfaces for Next Generation of Adaptive Computing Systems

DESCOMP: A New Design Space Exploration Approach

Design Space Navigation for Neighboring Power-Performance Efficient Microprocessor Configurations

An Efficient Frequency Scaling Approach for Energy-Aware Embedded Real-Time Systems

Pervasive Computing and Communication

Towards Autonomic Networking Using Overlay Routing Techniques

Context-Based Storage Management for Wearable and Portable Devices

A File System for System Programming in Ubiquitous Computing

Backmatter

Premium Partner