Xiayu Hua

Use Two-Level Rejuvenation to Combat Software Aging and Maximize Average Resource Performance

Software aging is a common phenomenon which is often manifested through system performance degradation. Rejuvenation is one of the most commonly used approaches to handle issues caused by software aging. To combat resource performance degradation and at the same time maintain maximized average resource performance, we present a two-level rejuvenation strategy, i.e., interleaving a set of n warm rejuvenations with one cold rejuvenation. Our target is to find the optimal n that maximizes system average performance. We first define a resource model that takes into consideration of performance degradation and two-level rejuvenations. Based on the resource model, we formally analyze the resource supply and present the MAX-PERFORMANCE algorithm to determine the optimal rejuvenation pattern that maximizes the average resource performance. The simulation results show that with a two-level rejuvenation strategy, we can achieve 25.22% higher average resource performance compared with a single level rejuvenation strategy.

Maximize System Reliability for Long Lasting and Continuous Applications

In this paper, we use software rejuvenation as a preventive and proactive fault-tolerance technique to maximize the level of reliability for continuous and safety critical systems. We take both transient faults caused by software aging effects and network transmission faults into consideration and mathematically analyze the optimal software rejuvenation period that maximizes system’s reliability. The theoretical result is verified through empirical studies.

BaaS (Bluetooth-as-a-sensor): conception, design and implementation on mobile platforms

As network connectivity becomes more capable, mobile devices are evolving into sensory data accumulators. Bluetooth (BT) components, which are widely used for communication purposes, also have the potential to become contextual sensors by constantly listening to information broadcast by nearby Bluetooth Low Energy (BLE) beacons. Compared to traditional Micro-Electro-Mechanical (MEMs) based contextual sensors, Bluetooth-as-a-Sensor (BaaS) provides a wider sensing spectrum and more comprehensive environmental information. However, current implementations of BT are optimized as a data transmitter, therefore deploying BaaS on a traditional mobile platform would cause an unacceptably high current drain and hence a significant reduction in battery life. Our objective is to conquer the current drain problem associated with having continuous wireless BT sensing. We provide a novel BaaS-based architecture which utilizes an energy-efficient sensor fusion core (SFC) to execute heavy-duty and long-standing tasks. We also present an optimized duty cycle algorithm that minimizes the duty cycle while guaranteeing an applications QoS requirements. Both BaaS architecture and algorithm are implemented and deployed on a Moto X platform and then applied to an indoor location service for consumer use validation. The performance of the BaaS-based architecture is evaluated for both average current drain and location accuracy. Data measured on Moto X shows that when using the BaaS architecture, the battery life is 5 times longer than using the traditional BT architecture.

Reliability guaranteed energy minimization on mixed-criticality systems

Analyze resource demand of MC task set under reliability and deadline constraints.Develop a heuristic approach to solve the formulated problem.Evaluate the proposed approach through simulation under various scenarios.Achieve up to 10% more energy saving comparing with the existing approaches. This paper studies the energy minimization problem in mixed-criticality systems that have stringent reliability and deadline constraints. We first analyze the resource demand of a mixed-criticality task set that has both reliability and deadline requirements. Based on the analysis, we present a heuristic task scheduling algorithm that minimizes systems energy consumption and at the same time also guarantees systems reliability and deadline constraints. Extensive experiments are conducted to evaluate and validate the performance of the proposed algorithm. The empirical results show that the algorithm further improves energy saving by up to 10% compared with the approaches proposed in our earlier work.

Schedulability Analysis for Real-Time Task Set on Resource with Performance Degradation and Periodic Rejuvenation

Most schedulability analyses in the literature assume that the performance of computing resource does not change over time. However, due to ever increased complexity of computer systems, software aging issues become more difficult, if not impossible, to eradicate. Hence, the assumption that computing resource has a constant performance in its entire lifetime does not hold in real world long-standing systems. In this paper, we study real-time task schedulability under a resource model that the resources performance degrades with a known degradation function and the resource is periodically rejuvenated. The resource model is referred to as P2-resource model for performance degradation and periodic rejuvenation. We address three real-task schedulability related questions under the P2-resource model, i.e., (1) resource supply bounds of the P2-resource, (2) task set utilization bounds under Earliest Deadline First (EDF) and Rate Monotonic (RM) scheduling policies, respectively, and (3) experimentally study the tightness of the bounds developed, and the impact of resource degradation rate, rejuvenation period, and rejuvenation cost on the bounds.

Periodic resource integration

Developed an integration method for two arbitrary fixed-pattern periodic resources.Provided the lower and upper bounds of the available time of an integrated resource.Provided a theoretical schedulability analysis for an integrated periodic resource.Provided an empirical study of task schedulability on an integrated resource. Scheduling periodic real-time tasks on multiple periodic resources is an emerging research issue in the real-time scheduling community and has drawn increased attention over the last few years. This paper studies a sub-category of the scheduling problem which focuses on scheduling a periodic task on multiple periodic resources where none of these resources have sufficient capacity to support the task. Instead of splitting the task into sub-tasks, which is not always practical in real systems, we integrate resources together to jointly support the task. First, we develop a method to integrate two fixed but arbitrary pattern periodic resources into an equivalent periodic resource. Second, for two periodic resources with unknown but fixed resource occurrence patterns, we give the lower and upper bounds of the available time provided by an integrated periodic resource within a period. Third, we present theoretical and empirical analysis on the schedulability of a non-splittable periodic task on two periodic resources and their integrated periodic resource.

Enhancing Throughput of Hadoop Distributed File System for Interaction-Intensive Tasks

The performance of the Hadoop Distributed File System (HDFS)decreases dramatically when handling interaction-intensive files, i.e., files that have relatively small size but are accessed frequently. The paper analyzes the cause of throughput degradation issue when accessing interaction-intensive files and presents an enhanced HDFS architecture along with an associated storage allocation algorithm that overcomes the performance degradation problem. Experiments have shown that with the proposed architecture together with the associated storage allocation algorithm, the HDFS throughput for interaction-intensive files increase 300% in average with only a negligible performance decrease for large data set tasks.

Implementing UHF RFID Reader on Smartphone Platform for IoT Sensing

As a core component of the Internet of Things technology (IoT), Radio Frequency Identification (RFID) tagged items will add billions, perhaps trillions, of objects to the Internet. As a result, uses of Ultra High Frequency (UHF) RFID sensing become massive ranging from logistics, retail and healthcare to homes and even entire smart cities. Under this trend, mobile UHF RFID scanners also need to evolve. Consumers will interact with their surroundings via tagged RFID items taking full advantage of the advancing IoT. For mainstream consumer smartphones, unfortunately, UHF RFID connectivity has yet to be fully integrated. The major challenges are: 1) the compatibility of an RFID reader module to the host platform, 2) Radio Frequency (RF) signal coexistence interference between the RFID reader and other sensor/RF technologies, and 3) the unacceptable high current drain caused by RFID active scanning. In this paper, we present a design and implementation of a novel modular UHF RFID scanning subsystem, the UHF RFID reader module, on a Motorola Moto-Z smartphone. This module is fully integrated with an Android 7.0 Operating System (OS) and directly interconnects with the low-level smartphone hardware and software framework. With the new antenna design and the signal spectrum analysis, we guarantee the RF isolation of the Mod with the smartphone’s other native wireless components and sensors. Our design and implementation also address the current drain issue and extends the battery life of Moto-Z smartphone up to 30.4 hours with IoT RFID scanning.