We characterize high-performance streaming applications as a new and distinct domain of programs that is becoming increasingly important.

An integrated platform which is capable of meeting the requirements of both traditional real-time control processing and multimedia processing has enormous potential for accommodating various kinds of new applications. However, except for the simplest of situations, few, if any, research or commercial systems successfully provide architectural and OS mechanisms which can efficiently support both hard real-time computation and multimedia soft real-time computation. In this paper, we propose a multimedia server executing on multiprocessor real-time operating systems to provide different classes of guarantee to support both types of processing. The multimedia server supports multiple periodic multimedia streams with a capability for graceful QoS degradation during system overload. In this paper we (i) discuss realistic system implementation issues on the SGI IRIX/REACT/PRO operating system, (ii) develop several multimedia server scheduling algorithms, and (iii) present a performance evaluation. We chose the SGI system as an implementation platform because it is being used more and more for multimedia applications. Our performance evaluation demonstrates that a multimedia server algorithm based on a flexible, proportional allocation scheme provides the best performance and that simple iterative scheduling is adequate for handling graceful degradation of the multimedia streams. We consider issues such as server size and period as well as the impact of context switch overhead on the performance. We also show that for applications which require integrated resource sharing, neither the frame scheduler nor the deadline scheduler supplied in the IRIX/REACT/PRO OS

Streaming programs represent an increasingly important and widespread class of applications that holds unprecedented opportunities for high-impact compiler technology. Unlike sequential programs with obscured dependence information and complex communication patterns, a stream program is naturally written as a set of concurrent filters with regular steady-state communication. The StreamIt language aims to provide a natural, high-level syntax that improves programmer productivity in the streaming domain. At the same time, the language imposes a hierarchical structure on the stream graph that enables novel representations and optimizations within the StreamIt compiler. We define the "stream dependence function", a fundamental relationship between the input channels of two filters in a stream graph. We also describe a suite of stream optimizations, a denotational semantics for validating these optimizations, and a novel phased scheduling algorithm for stream graphs. In addition, we have implemented a prototype of the StreamIt optimizing compiler that is showing promising results.

Software Defined Radios are becoming more and more prevalent. Especially in the radio amateur community, Software Defined Radios are a big success. The wireless industry also has considerable interest in the dynamic reconfigurability and other advantages of Software Defined Radios. Our research focuses on the latency of Software Defined Radios and its impact on throughput in modern wireless protocols. Software Defined Radio systems often employ a bus system to transfer the samples from a radio frontend to the processor which introduces a non-negligible latency. Additionally, the signal processing calculations on general-purpose processors introduce additional latencies that are not found on conventional radios. This work concentrates on one particular Software Defined Radio system called GNU Radio, an open source Software Defined Radio application, and one of its hardware components, the Universal Software Radio Peripheral (USRP), and analyzes its receive and transmit latencies. We will use these measurements to characterize the performance impact on IEEE 802.15.4 implementation in GNU Radio. Additionally, we present two Software Defined Radio implementations of short-range radio standards, a FSK scheme used in the Chipcon CC1000 radio, and the physical layer of IEEE 802.15.4. We use these implementations for round trip time measurements and introduce two sample applications, a physical layer bridge between the FSK scheme and IEEE 802.15.4, and a dual channel receiver that receives two radio channels concurrently.

An integrated platform which is capable of meeting requirements of both traditional real-time control processing and multimedia processing has enormous potential for accommodating various kinds of new applications. However, few, if any, research or commercial systems successfully provide architectural and OS mechanisms which can efficiently support both deterministic hard real-time computation and less deterministic multimedia soft real-time computation. In this paper, we propose a multimedia server executing on the Spring real-time system to provide different classes of guarantee to support both types of processing. Using a planning based scheduler, the Spring system is a multiprocessor real-time system which was developed to satisfy the requirements of deterministic guarantees for hard real-time tasks. Adding a multimedia server enables Spring to also support multiple periodic multimedia streams with a capability for QOS graceful degradation of the multimedia streams during system ov...