Mission-critical embedded systems simultaneously run multiple graphics-processing-unit (GPU) computing tasks with different criticality and timeliness requirements. Considerable research effort has been dedicated to supporting the preemptive priority scheduling of GPU kernels. However, hardware-supported preemption leads to lengthy scheduling delays and complicated designs, and most software approaches depend on the voluntary yielding of GPU resources from restructured kernels. We propose a preemptive GPU kernel scheduling scheme that harnesses the idempotence property of kernels. The proposed scheme distinguishes idempotent kernels through static source code analysis. If a kernel is not idempotent, then GPU kernels are transactionized at the operating system (OS) level. Both idempotent and transactionized kernels can be aborted at any point during their execution and rolled back to their initial state for reexecution. Therefore, low-priority kernel instances can be preempted for high-priority kernel instances and reexecuted after the GPU becomes available again. Our evaluation using the Rodinia benchmark suite showed that the proposed approach limits the preemption delay to 18 ms in the 99.9th percentile, with an average delay in execution time of less than 10 percent for high-priority tasks under a heavy load in most cases.

Digital video is prominent big data spread all over the Internet. It is large not only in size but also in required processing power to extract useful information. Fast processing of excessive video reels is essential on criminal investigations, such as terrorism. This demo presents an extensible video processing framework in Apache Hadoop to parallelize video processing tasks in a cloud environment. Except for video transcending systems, there have been few systems that can perform various video processing in cloud computing environments. The framework employs FFmpeg for a video coder, and OpenCV for a image processing engine. To optimize the performance, it exploits MapReduce implementation details to minimize video image copy. Moreover, FFmpeg source code was modified and extended, to access and exchange essential data and information with Hadoop, effectively. A face tracking system was implemented on top of the framework for the demo, which traces the continuous face movements in a sequence of video frames. Since the system provides a web-based interface, people can try the system on site. In an 8-core environment with two quad-core systems, the system shows 75% of scalability.

The Physically-Asynchronous Logically-Synchronous (PALS) system is a recently proposed architectural pattern for cyber-physical systems. It guarantees a logically synchronous design abstraction for real-time distributed computations. In this work, we develop a new middleware, called PALSware, to support an efficient and robust implementation of the PALS system and its extensions. PALSware guarantees consistency in distributed applications by eliminating any asynchronous interactions resulting from distributed clocks and node failures. We present a layered design for this middle-ware that is both reusable in different system architectures and can be extended with architecture-specific solutions for fault management. We demonstrate the middleware for an academic control testbed and show the consistency in a fault injection framework designed for this middleware.

This demonstration presents our NASS (Network Aware Supervisory System) framework prototype for medical device integration systems. The NASS framework interconnects medical devices over wireless for convenience, seamlessness and sanitation, and provides safety-guaranteed supervision. Our prototype was developed in Sun Java Real-time Environment. Real-time Java provides well-formed convenience of dynamically loading and unloading medical application logic and safety rules on the fly in real-time environments. To tackle the complexity of using real-time Java in the safety-critical system, we also applied HW/SW codesign method. Real-time Java Environment + Linux operating system may not be robust enough for medical devices to fully rely on. In our prototype, the supervisor software in Java performs all logical decisions including contingency plan generation derived from the safety rules. Once logic is decided, the decisions and plans for the devices are delivered to the hardware implemented in FPGA at each device to physically drive medical equipments. Since the execution of decisions and plans are delegated to the hardware, any failure in software does not harm the integrated safety. Our demonstration shows how safety is managed in different kinds of failures from wireless network failures to device software failures.

Most dynamic voltage and frequency scaling (DVS) techniques adjust only CPU parameters, however, recent embedded systems provide multiple adjustable clocks which can be independently tuned. When considering multiple components, energy optimal frequencies depend on task set characteristics such as the number of CPU and memory access cycles. In this work, we propose a realistic energy model considering multiple components with individually adjustable frequencies such as CPU, system bus and memory, and related task set characteristics. The model is validated on a real platform and shows less than 2% relative error compared to measured values. Based on the proposed energy model, we present an optimal static frequency assignment scheme for multiple DVS components to schedule a set of periodic realtime tasks. We simulate the energy gain of the proposed scheme compared to other DVS schemes for various task and system configurations, showing up to a 20% energy reduction.

BCMCS (broadcast and multicast services) greatly increase the capacity of CDMA2000 mobile networks for multimedia broadcasting, but only require minor changes to existing radio and core network protocols. To provide high-quality multimedia services over error-prone wireless networks, we need to control delay, because it is one of the most important QoS parameters for multimedia applications. We identified error control using Reed-Solomon (RS) forward error correction (FEC) in the medium access control (MAC) layer as a significant cause of delay variation, and have proposed a novel analytic model that can predict the time required for RS decoding at a mobile, under different levels of block interleaving and for varying conditions of the Rayleigh fading channel at that mobile. This model enables us to estimate the buffer size required to ensure seamless multimedia services.

Superposition interference cancellation (SIC) has been highlighted due to its high spatial reuse efficiency in wireless networks. It exploits superposition aspect of multiple signals to cancel interferences. We present three basic SIC schemes, which are (1) superposition demodulation scheme, (2) superposition decoding scheme and (3) iterative superposition decoding scheme in this paper. When selecting an SIC algorithm among them, the computational overhead and cancellation performance have a trade-off relationship. In a fair communication environment, the receiver can save energy with little loss of cancellation performance by adopting a simple SIC algorithm, such as superposition demodulation. If the network condition looks bad, the node should apply a sophisticated SIC scheme, like iterative superposition decoding, to accommodate near-optimal performance of cancellation. Selecting a different SIC scheme in terms of network condition is called adaptive interference cancellation. The adaptive cancellation scheme draws fairly good interference cancellation performance with relatively low mean computational overhead to save energy consumption.

A new generation of wireless personal area networks (WPANs) is intended for high data rate and multimedia applications. The multiband OFDM alliance (MBOA) is currently developing a new PHY and MAC protocol in WPANs, which fit the needs of this mass market. The purpose of this paper is the power efficient multimedia multicast protocol to apply MBOA specification. In general, multicast traffic is transmitted at lowest data rate without retransmission. If the receivers are in near position from sender, transmission delay is increased and it leads to transmission/receiving power wastage. In this paper, we propose the selective multimedia multicast protocol (SMMP) that tries to reduce the power by separation scheme using punctured convolutional code and concatenation scheme. In addition, this protocol provides reliability through several feedback information of receivers. From our simulation study, we show that the proposed mechanism reduces the power consumption and achieves improved performance.