Detailed degradation information can help uncover the underlying mechanisms of system degradation and provide essential guidance for accurate prognostics. However, obtaining such information in practice is often costly and challenging, which has motivated the development of prognostics methods that integrate fault diagnosis. Most existing methods, however, primarily focus on degradation magnitude while neglecting trend information, and they also lack sufficient robustness to uncertainties. As a result, they often suffer from low prediction accuracy and limited prediction horizons. To overcome these challenges, this paper proposes a novel diagnosis-integrated prognostics framework in which the remaining useful life prediction task is formulated as an explicit parameter optimization problem. Specifically, the robustness of optimization algorithms contributes to reliable prediction accuracy even under imprecise prior knowledge and low-quality data, while the integration of degradation trend information enhances the ability to capture long-term dynamics, which in turn extends the prediction horizon. To better reflect realistic operating conditions, concurrent actuator and sensor faults are considered. The approach first extracts both magnitude and trend information via fault diagnosis, and then employs parameter optimization for RUL prediction. The effectiveness of the proposed method is demonstrated through simulation studies of actuator and sensor degradation in a grain processing system.

In today's computer systems, the disk I/O subsystem is often identified as the major bottleneck to system performance. One proposed solution is the so called redundant array of inexpensive disks (RAID). We examine the performance of two of the most promising RAID architectures, the mirrored array and the rotated parity array. First, we propose several scheduling policies for the mirrored array and a new data layout, group-rotate declustering, and compare their performance with each other and in combination with other data layout schemes. We observe that a policy that routes reads to the disk with the smallest number of requests provides the best performance, especially when the load on the I/O system is high. Second, through a combination of simulation and analysis, we compare the performance of this mirrored array architecture to the rotated parity array architecture. This latter study shows that: 1) given the same storage capacity (approximately double the number of disks), the mirrored array considerably outperforms the rotated parity array; and 2) given the same number of disks, the mirrored array still outperforms the rotated parity array in most cases, even for applications where I/O requests are for large amounts of data. The only exception occurs when the I/O size is very large; most of the requests are writes, and most of these writes perform full stripe write operations.

Disk mirroring has found widespread use in computer systems as a method for providing fault tolerance. In addition to increasing reliability, a mirrored disk can also reduce I/O response time by supporting the execution of parallel I/O requests. The improvement in I/O efficiency is extremely important in a real-time system, where each computational entity carries a deadline. In this paper, we present two classes of real-time disk scheduling policies, RT-DMQ and RT-CMQ, for a mirrored disk I/O subsystem and examine their performance in an integrated real-time transaction system. The real-time transaction system model is validated on a real-time database testbed, called RT-CARAT. The performance results show that a mirrored disk I/O subsystem can decrease the fraction of transactions that miss their deadlines over a single disk system by 68%. Our results also reveal the importance of real-time scheduling policies, which can lead up to a 17% performance improvement over non-real-time policies in terms of minimizing the transaction loss ratio.