Wanghong Yuan

Energy-efficient soft real-time CPU scheduling for mobile multimedia systems

This paper presents GRACE-OS, an energy-efficient soft real-time CPU scheduler for mobile devices that primarily run multimedia applications. The major goal of GRACE-OS is to support application quality of service and save energy. To achieve this goal, GRACE-OS integrates dynamic voltage scaling into soft real-time scheduling and decides how fast to execute applications in addition to when and how long to execute them. GRACE-OS makes such scheduling decisions based on the probability distribution of application cycle demands, and obtains the demand distribution via online profiling and estimation. We have implemented GRACE-OS in the Linux kernel and evaluated it on an HP laptop with a variable-speed CPU and multimedia codecs. Our experimental results show that (1) the demand distribution of the studied codecs is stable or changes smoothly. This stability implies that it is feasible to perform stochastic scheduling and voltage scaling with low overhead; (2) GRACE-OS delivers soft performance guarantees by bounding the deadline miss ratio under application-specific requirements; and (3) GRACE-OS reduces CPU idle time and spends more busy time in lower-power speeds. Our measurement indicates that compared to deterministic scheduling and voltage scaling, GRACE-OS saves energy by 7% to 72% while delivering statistical performance guarantees.

Design and evaluation of a cross-layer adaptation framework for mobile multimedia systems

Mobile multimedia systems must provide application quality of service (QoS) in the presence of dynamically varying and multiple resource constraints (e.g., variations in available CPU time, energy, and bandwidth). Researchers have therefore proposed adaptive systems that can respond to changing resource availability and application demands. All system layers can benefit from adaptation, but fully exploiting these benefits requires a new cross-layer adaptation framework to coordinate the adaptations in the different layers. This paper presents such a framework and its first prototype, called GRACE-1. The framework supports application QoS under CPU and energy constraints via coordinated adaptation in the hardware, OS, and application layers. Specifically, GRACE-1 uses global adaptation to handle large and long-term variations, setting application QoS, CPU allocation, and CPU frequency/voltage to qualitatively new levels. In response to small and temporary variations, it uses local adaptation within each layer. We have implemented the GRACE-1 prototype on an HP laptop with an adaptive processor. Our experimental results show that, compared to previous approaches that exploit adaptation in only some of the layers or in an uncoordinated way, GRACE-1 can provide higher overall system utility in several cases.

An XML-based Quality of Service Enabling Language for the Web

In this paper, we introduce an XML-based hierarchical QoS markup language, called HQML, to enhance distributed multimedia applications on the World Wide Web (WWW) with quality of service (QoS) capability. The design ofHQML is based on two observations: (1) the absence of a systematic QoS specification language, that can be used by distributed multimedia applications on the WWW to utilize the state-of-the-art QoS management technology and (2) the power and popularity of XML to deliver richly structured contents over the Web. HQML allows distributed multimedia applications to specify all kinds of application-specific QoS policies and requirements. During runtime, the HQML Executor translates the HQML file into desired data structures and cooperates with the QoS proxies that assist applications in end-to-end QoS negotiation, setup and enforcement. In order to make QoS services tailored toward user preferences and meet the challenges of uncertainty in the distributed heterogeneous environments, the design of HQML is featured as interactive andflexible . In order to allow application developers to create HQML specifications correctly and easily, we have designed and developed a unified visual QoS programming environment, called QoSTalk. In QoSTalk, we adopt a grammatical approach to perform consistency check on the visual QoS specifications and generate HQML files automatically. Finally, we introduce the distributed QoS compiler, which performs the automatic mappings between application- and resource-level QoS parameters to relieve the application developer of the burden of dealing with low-level QoS specifications.

Energy-efficient CPU scheduling for multimedia applications

This article presents the design, implementation, and evaluation of EScheduler, an energy-efficient soft real-time CPU scheduler for multimedia applications running on a mobile device. EScheduler seeks to minimize the total energy consumed by the device while meeting multimedia timing requirements. To achieve this goal, EScheduler integrates dynamic voltage scaling into the traditional soft real-time CPU scheduling: It decides at what CPU speed to execute applications in addition to when to execute what applications. EScheduler makes these scheduling decisions based on the probability distribution of cycle demand of multimedia applications and obtains their demand distribution via online profiling.We have implemented EScheduler in the Linux kernel and evaluated it on a laptop with a variable-speed CPU and typical multimedia codecs. Our experimental results show four findings: first, the cycle demand distribution of our studied codecs is stable or changes slowly. This stability implies the feasibility to perform our proposed energy-efficient scheduling with low overhead. Second, EScheduler delivers soft performance guarantees to these codecs by bounding their deadline miss ratio under the application-specific performance requirements. Third, EScheduler reduces the total energy of the laptop by 14.4% to 37.2% relative to the scheduling algorithm without voltage scaling and by 2% to 10.5% relative to voltage scaling algorithms without considering the demand distribution. Finally, EScheduler saves energy by 2% to 5% by explicitly considering the discrete CPU speeds and the corresponding total power of the whole laptop, rather than assuming continuous speeds and cubic speed-power relationship.

Practical voltage scaling for mobile multimedia devices

This paper presents the design, implementation, and evaluation of a <i>practical</i> voltage scaling (PDVS) algorithm for mobile devices primarily running multimedia applications. PDVS seeks to minimize the total energy of the whole device while meeting multimedia timing requirements. To do this, PDVS extends traditional real-time scheduling by deciding <i>what execution speed</i> in addition to when to execute what applications. PDVS makes these decisions based on the discrete speed levels of the CPU, the total power of the device at different speeds, and the probability distribution of CPU demand of multimedia applications. We have implemented PDVS in the Linux kernel and evaluated it on an HP laptop. Our experimental results show that PDVS saves energy substantially without affecting multimedia performance. It saves energy by 14.4% to 37.2% compared to scheduling algorithms without voltage scaling and by up to 10.4% compared to previous voltage scaling algorithms that assume an ideal CPU with continuous speeds and cubic power-speed relationship.

Quality of protection for mobile multimedia applications

In traditional computer systems, security is typically provided in a one-or-nothing manner; the system is either secure or insecure. Such an approach is insufficient for pervasive environments that contain heterogenous devices with varying computing resources. The small, portable handheld devices are often left unsecured due to their limited computing power. The approach is also inadequate for multimedia applications that require security as a controllable service attribute to maintain performance quality of service to levels that are acceptable to the users. Hence, we need a tunable and differentiable security framework. In this paper, we present a quality of protection (QoP) framework that resolves the inadequacies of the one-or-nothing approach by providing differential security levels for different device, user and application security requirements and preferences. We show that our QoP framework is necessary for multimedia applications to achieve the best possible security and performance levels in pervasive environments.

GRACE-2: integrating fine-grained application adaptation with global adaptation for saving energy

Energy efficiency has become a primary design criterion for mobile multimedia devices. Prior work has proposed saving energy through coordinated adaptation in multiple system layers, in response to changing application demands and system resources. The scope and frequency of adaptation pose a fundamental conflict in such systems. The Illinois GRACE project addresses this conflict through a hierarchical solution which combines: 1) infrequent (expensive) global adaptation that optimises energy for all applications in the system, 2) frequent (cheap) per-application (or per-app) adaptation that optimises for a single application at a time. This paper demonstrates the benefits of the hierarchical adaptation through a second-generation prototype, GRACE-2. Specifically, it shows that in a network bandwidth constrained environment, per-app application adaptation yields significant energy benefits over and above global adaptation.

Integration of dynamic voltage scaling and soft real-time scheduling for open mobile systems

Battery-powered mobile devices are becoming increasingly important computing platforms, which require low energy consumption while meeting the resource demands of a dynamic application workload. Most proposed dynamic voltage scaling (DVS) algorithms, targeting either best-effort or hard real-time systems, however, cannot be directly applied to such open mobile systems. This paper presents a framework to integrate DVS into soft real-time (SRT) scheduling for open mobile systems, achieving energy saving of DVS while preserving resource guarantees of SRT scheduling. The integrated framework makes three major contributions. First, multimedia applications reserve resource based on their average resource usage, without the knowledge of worst-case execution time, which is difficult to estimate in an open mobile environment. Second, the SRT scheduling ensures the correctness of reservation admission and enforcement in a variable speed context. Finally, the DVS manager reduces the processor energy consumption by utilizing the unallocated resource, reclaiming the allocated but unused resource, or avoiding the unused resource. Our extensive simulation results demonstrate that our framework is able to save 4% to 32% energy while slightly affecting application performance.

R-EDF: a reservation-based EDF scheduling algorithm for multiple multimedia task classes

Multimedia applications became an integral part of the ubiquitous computing environment in general-purpose systems. However, current general-purpose operating systems cannot effectively support their soft real-time requirements. This paper presents a reservation-based preemptive earliest deadline first (R-EDF) algorithm to provide a predictable scheduling framework in an open shared environment. The major contributions of R-EDF are: it delivers timing guarantees to four soft real-time task classes such as period class with constant and various processing time, aperiodic class and event-based class; it provides starvation protection for best-effort tasks; it provides overrun protection and handling so that an overrun task will not cause other tasks to miss deadlines and will finish its work within an acceptable and predictable time bound; and it allows optimistic reservation and incurs low overhead cost for reservation and admission control. Our experimental results show the low overhead and correctness of R-EDF even when there is resource contention.

A middleware framework coordinating processor/power resource management for multimedia applications

It is challenging to reduce the processor power consumption while meeting the processor resource requirement of distributed multimedia applications in portable computers. This paper presents a middleware framework coordinating the processor/power resource management (PPRM) in a mobile computing environment. Our framework has four major contributions: (1) providing a power-aware resource reservation mechanism, where admission control is based on the processor utilization and power availability; (2) adjusting the speed and corresponding power consumption of the processor depending upon events and triggered by the change of the system workload or power availability; (3) updating reservation contracts of multimedia applications to maintain their resource requirements while adjusting the processor speed; (4) notifying applications about the change of resource status to enable them to adapt their behavior and complete tasks before the power drains. Our experimental results show the effectiveness of the coordinating PPRM framework to save energy and maintain the resource requirements of multimedia applications.