Failure Logs Research Articles

Long term risk of heart failure in adult cancer survivors: a systematic review and meta-analysis

Abstract Background Evaluation of the cardiotoxic effects from cancer is most commonly evaluated at the time of chemotherapy. Although cancer survivorship populations are rapidly expanding, there is a paucity of evidence regarding the frequency and causes of heart failure (HF) >5 years after cardiotoxic therapy. Objectives This systematic review sought to investigate the relationship between cardiotoxic cancer therapies and late-onset HF. Methods We searched the EMBASE and MEDLINE databases for studies reporting HF in adult survivors (≥50 years old), who were ≥5 years post-breast/lymphoma cancer therapy. A random effects model was used to examine the associations of HF. A meta-regression analysis was implemented to determine causes of heterogeneity. Results Thirteen papers were included in the review and meta-analysis (Figure 1), comprising 190,259 participants (mean age 53.5 years, 93% female). The risk of HF was increased (overall log risk-ratio 0.39 (95%CI (0.16-0.62)), (Figure 2). Cardiotoxic treatment, compared with cancer alone, provided a similar risk (log risk ratio of 0.38 [95% CI 0.02-0.77]). In the breast cancer population ratio (11 studies) the overall heart failure log risk was 0.41 [95% CI 0.13-0.69]). Although heterogeneity was significant (I²=77.17), this was explained by differences in patient characteristics; once multivariable analysis accounted for follow up (0.99, 95%CI (0.97-0.99), p=0.047), age (1.14, 95%CI (1.04-1.25), p=0.003) and hypertension (0.95, 95%CI (0.92-0.98), p<0.001), residual heterogeneity was low (I²=28.73). Conclusions HF is increased in adult cancer survivors, associated with cardiotoxic cancer therapy and standard risk factors. Consideration should be given for selective screening for HF in this at-risk population.Fig 1:Literature Search Flow DiagramFig 2:Incident HF in Cancer Survivors

UniLoc: Unified Fault Localization of Continuous Integration Failures

Continuous integration (CI) practices encourage developers to frequently integrate code into a shared repository. Each integration is validated by automatic build and testing such that errors are revealed as early as possible. When CI failures or integration errors are reported, existing techniques are insufficient to automatically locate the root causes for two reasons. First, a CI failure may be triggered by faults in source code and/or build scripts, whereas current approaches consider only source code. Second, a tentative integration can fail because of build failures and/or test failures, whereas existing tools focus on test failures only. This article presents UniLoc, the first unified technique to localize faults in both source code and build scripts given a CI failure log, without assuming the failure’s location (source code or build scripts) and nature (a test failure or not). Adopting the information retrieval (IR) strategy, UniLoc locates buggy files by treating source code and build scripts as documents to search and by considering build logs as search queries. However, instead of naïvely applying an off-the-shelf IR technique to these software artifacts, for more accurate fault localization, UniLoc applies various domain-specific heuristics to optimize the search queries, search space, and ranking formulas. To evaluate UniLoc, we gathered 700 CI failure fixes in 72 open source projects that are built with Gradle. UniLoc could effectively locate bugs with the average mean reciprocal rank value as 0.49, mean average precision value as 0.36, and normalized discounted cumulative gain value as 0.54. UniLoc outperformed the state-of-the-art IR-based tool BLUiR and Locus. UniLoc has the potential to help developers diagnose root causes for CI failures more accurately and efficiently.

Long-term immunosuppressive treatment is not associated with worse outcome in patients hospitalized in the intensive care unit for septic shock: the PACIFIC study

BackgroundExcept in a few retrospective studies mainly including patients under chemotherapy, information regarding the impact of immunosuppressive therapy on the prognosis of patients admitted to the intensive care unit (ICU) for septic shock is scarce. Accordingly, the PACIFIC study aimed to asses if immunosuppressive therapy is associated with an increased mortality in patients admitted to the ICU for septic shock.MethodsThis was a retrospective epidemiological multicentre study. Eight high enroller centres in septic shock randomised controlled trials (RCTs) participated in the study. Patients in the “exposed” group were selected from the screen failure logs of seven recent RCTs and excluded because of immunosuppressive treatment. The “non-exposed” patients were those included in the placebo arm of the same RCTs. A multivariate logistic regression model was used to estimate the risk of death.ResultsAmong the 433 patients enrolled, 103 were included in the “exposed” group and 330 in the “non-exposed” group. Reason for immunosuppressive therapy included organ transplantation (n = 45 [44%]) or systemic disease (n = 58 [56%]). ICU mortality rate was 24% in the “exposed” group and 25% in the “non-exposed” group (p = 0.9). Neither in univariate nor in multivariate analysis immunosuppressive therapy was associated with a higher ICU mortality (OR: 0.95; [95% CI 0.56–1.58]: p = 0.86 and 1.13 [95% CI 0.61–2.05]: p = 0.69, respectively) or 3-month mortality (OR: 1.13; [95% CI 0.69–1.82]: p = 0.62 and OR: 1.36 [95% CI 0.78–2.37]: p = 0.28, respectively).ConclusionsIn this study, long-term immunosuppressive therapy excluding chemotherapy was not associated with significantly higher or lower ICU and 3-month mortality in patients admitted to the ICU for septic shock.

Software Failure Log Analysis for Engineers—Review

The use of automated methods for log analysis is unavoidable in any large company; therefore, it has attracted attention from engineers and researchers. As a result, the number of articles in the field grows yearly and new approaches are frequently proposed. Unfortunately, published research works only sometimes meet the needs of engineers wishing to apply the methods in real-life systems. A common issue is that the method’s benefits often do not compensate for the effort required for its implementation and maintenance. Therefore, engineers must understand the pros and cons of full-scale applications, including the implementation details and the required effort. This work provides a comprehensive review of automated log analysis methods and aims to offer a guide for software engineers who fix integration and production failures. The article categorizes and provides an overview of existing methods and assesses their implementation and maintenance costs, as well as the feasibility of the methods. The article also identifies and describes the shortcomings of existing methods, including concept drift, which is not addressed with sufficient attention, as well as the lack of online benchmarks and the interpretation of the log sequence as a language, without an in-depth analysis of its properties. Despite growing efforts to provide feasible and widely adopted solutions, many reference implementations are unavailable. Consequently, the time and computation complexities differ between various implementations of the same approaches, making the results of research work difficult to replicate in real-life scenarios.

Acquisition and Processing of UAV Fault Data Based on Time Line Modeling Method

The number of Unmanned Aerial Vehicles (UAVs) used in various industries has increased exponentially, and abnormal detection of UAVs is one of the primary technical means to ensure that UAVs can work normally. Currently, most anomaly detection models are trained using on-board logs from drones. However, in some cases, using these logs can be problematic due to data encryption, inconsistent descriptions of characteristics, and imbalanced positive and negative samples. Consequently, the on-board logs of UAVs may not be directly usable for training anomaly detection models. Given the above problems, this paper proposes a Time Line Modeling (TLM) method based on the UAV software-in-the-loop (SITL) simulation environment to obtain and process the on-board failure logs of drones. The Time Line Modeling method includes two stages: the Fault Time Point Anchoring Method and Fault Time Window Stretching Method. First, based on the SITL simulation environment, multiple flight missions were constructed. Failures of several common components of UAVs are designed. Secondly, the fault’s initial location and end location are determined by the method of Fault Time Point Anchoring, and the original collection of tagged UAV’s on-board data is realized. Then, in terms of data processing, the features that are not universal are removed, and the flight data of the UAV is optimized by using the data balance method of Time Window Stretching to achieve the balance of normal data and abnormal data. Finally, use of algorithms such as Sequential Minimal Optimization (SMO), Random Forest (RF), and Convolutional Neural Network (CNN) were used to experiment with the processed data. The experimental results showed that the data set obtained based on this method can be effectively applied to the training of machine learning-based anomaly detection models.

Run-time failure detection via non-intrusive event analysis in a large-scale cloud computing platform

Cloud computing systems fail in complex and unforeseen ways due to unexpected combinations of events and interactions among hardware and software components. These failures are especially problematic when they are silent, i.e., not accompanied by any explicit failure notification, hindering the timely detection and recovery.In this work, we propose an approach to run-time failure detection tailored for monitoring multi-tenant and concurrent cloud computing systems. The approach uses a non-intrusive form of event tracing, without manual changes to the system’s internals to propagate session identifiers (IDs), and builds a set of lightweight monitoring rules from fault-free executions. We evaluated the effectiveness of the approach in detecting failures in the context of the OpenStack cloud computing platform, a complex and “off-the-shelf” distributed system, by executing a campaign of fault injection experiments in a multi-tenant scenario. Our experiments show that the approach detects the failure with an F1 score (0.85) and accuracy (0.77) higher than the ones provided by the OpenStack failure logging mechanisms (0.53 and 0.50) and two non-session-aware run-time verification approaches (both lower than 0.15). Moreover, the approach significantly decreases the average time to detect failures at run-time (∼114 seconds) compared to the OpenStack logging mechanisms.

A Study of Failure Recovery and Logging of High-Performance Parallel File Systems

Large-scale parallel file systems (PFSs) play an essential role in high-performance computing (HPC). However, despite their importance, their reliability is much less studied or understood compared with that of local storage systems or cloud storage systems. Recent failure incidents at real HPC centers have exposed the latent defects in PFS clusters as well as the urgent need for a systematic analysis. To address the challenge, we perform a study of the failure recovery and logging mechanisms of PFSs in this article. First, to trigger the failure recovery and logging operations of the target PFS, we introduce a black-box fault injection tool called PFault , which is transparent to PFSs and easy to deploy in practice. PFault emulates the failure state of individual storage nodes in the PFS based on a set of pre-defined fault models and enables examining the PFS behavior under fault systematically. Next, we apply PFault to study two widely used PFSs: Lustre and BeeGFS. Our analysis reveals the unique failure recovery and logging patterns of the target PFSs and identifies multiple cases where the PFSs are imperfect in terms of failure handling. For example, Lustre includes a recovery component called LFSCK to detect and fix PFS-level inconsistencies, but we find that LFSCK itself may hang or trigger kernel panics when scanning a corrupted Lustre. Even after the recovery attempt of LFSCK, the subsequent workloads applied to Lustre may still behave abnormally (e.g., hang or report I/O errors). Similar issues have also been observed in BeeGFS and its recovery component BeeGFS-FSCK. We analyze the root causes of the abnormal symptoms observed in depth, which has led to a new patch set to be merged into the coming Lustre release. In addition, we characterize the extensive logs generated in the experiments in detail and identify the unique patterns and limitations of PFSs in terms of failure logging. We hope this study and the resulting tool and dataset can facilitate follow-up research in the communities and help improve PFSs for reliable high-performance computing.

Classification and Prediction of Software Incidents Using Machine Learning Techniques

An incident, in the perception of information technology, is an event that is not part of a normal process and disrupts operational procedure. This research work particularly focuses on software failure incidents. In any operational environment, software failure can put the quality and performance of services at risk. Many efforts are made to overcome this incident of software failure and to restore normal service as soon as possible. The main contribution of this study is software failure incidents classification and prediction using machine learning. In this study, an active learning approach is used to selectively label those data which is considered to be more informative to build models. Firstly, the sample with the highest randomness (entropy) is selected for labeling. Secondly, to classify the labeled observation into either failure or no failure classes, a binary classifier is used that predicts the target class label as failure or not. For classification, Support Vector Machine is used as a main classifier to classify the data. We derived our prediction models from the failure log files collected from the ECLIPSE software repository.

Why Are Women Less Represented in Intracerebral Hemorrhage Trials?

Fewer women than men tend to be enrolled in clinical trials of intracerebral hemorrhage. It is unclear whether this reflects lower prevalence of intracerebral hemorrhage in women, selection bias, or poor recruitment efforts. We undertook this study to examine differences between men and women in the reasons for exclusion from the iDEF trial (Intracerebral Hemorrhage Deferoxamine). The screen failure log included 29 different reasons for exclusion. Chi-square statistics were used to evaluate the differences in reasons for exclusion between men and women. A total of 38.2% of participants in iDEF were women. Three thousand nine hundred eighty-two women (45.7%) and 4736 men (54.3%) were screen failures (P<0.0001). Similar proportions of women (1.28%) and men (1.73%) were excluded due to inability to obtain consent (P=0.1). Patients or families declined participation in 1.26% of women versus 1.31% of men (P=0.9). More women than men failed screening because of age>80 (22.40% versus 12.61%; adjusted P=0.0007) and preexisting do-not-resuscitate/do-not-intubate (3.69% versus 2.83%; adjusted P=0.067). Lower rates of women enrollment in the iDEF trial may be attributed to older age. Inability to obtain consent or declining participation was similar between women and men, arguing against selection bias. Our findings should be confirmed in other intracerebral hemorrhage trials to determine best strategies to improve women's representation in future trials.

Prognosis of Wind Turbine Gearbox Bearing Failures using SCADA and Modeled Data

&#x0D; &#x0D; &#x0D; Predictive maintenance and condition monitoring systems for wind turbines have seen increased adoption to minimize downtime, reducing operation and maintenance costs. On today’s wind power plants, the integrated supervisory control and data acquisition (SCADA) system provides low- frequency operational data that can be leveraged to quantify a wind turbine’s health. The aim of this study is to utilize machine-learning techniques to predict axial cracking failures in wind turbine gearbox bearings up to 1 month ahead of time. The failures are assumed to have occurred when the investigated bearing was replaced. While current SCADA systems show the overall condition of a wind turbine, often they do not allow for the investigation of specific gearbox bearings’ health. To enrich bearing fault signatures, additional data are computed through physics-based models using gearbox design information. Based on SCADA data, modeled data, and bearing failure log data from an actual wind plant, the performances of different machine-learning models on unseen data are then evaluated using industry-standard metrics such as precision, recall, and F1 score. Results show the overall system performance enhancement in predicting bearing failure when modeled data are included with SCADA data. The reduction in terms of false alarms is about 50%, and improvement in terms of precision and F1 score is about 33% and 12% respectively, based on the best modeling case in this study.&#x0D; &#x0D; &#x0D;

Calculation of an SPTA set using the Dialog computer simulation system (Part 1. General provisions for the calculation of an SPTA set)

The paper describes the design concept of the DIALOG simulation software suite (SSS) intended for calculating the dependability indicators of electronic systems (ES) of random configuration, as well as solving problems associated with assuring the functional dependability of such systems. The DIALOG SSS employs a specially designed DIALOG-SINTEZ technology that enables automatic synthesis of event simulation models in the form of programs in the selected programming language. In DIALOG SSS, the input data include: system composition in the form of a combination of conventional blocks; criteria of failure event occurrence and repairs; random value parameters (failures of system components in various modes of operation, SPTA requests, etc.); stages of system operation and types of repairs; list of calculated indicators. For the purpose of ensuring the required indicators, the simulation models undergo statistical testing under modified indicators of random values in each new test. Based on the accumulated results of all the performed tests the required indicators are calculated. The DIALOG SSS consists of four components: DIALOG-NRS is intended for the calculation of dependability indicators of non-repairable redundant systems; DIALOG-RRS is intended for the calculation of dependability indicators of repairable redundant systems, as well as the number of and cost of warranty repairs; DIALOG-ZIP-NS is intended for SPTA sets calculation for simple non-redundant systems; DIALOG-ZIP-RS is intended for SPTA sets calculation for any redundant systems. SPTA set calculation is normally done using standard procedures described in regulatory documents. In solving the direct problem of optimal SPTA set calculation, the input data includes the required value of one of the two SPTA sufficiency indicators (SI) and type of costs that are to be optimized (minimized) in order to achieve the target values of SI. In solving the inverse problem of optimal SPTA set calculation, it is required to ensure the specified costs of the initial SPTA supply. As the SPTA supply SI, the mean spare parts (SP) supply delay time out of SPTA Δtd.SPTA and SPTA availability coefficient Ca.SPTA are used. SPTA optimization using the DIALOG-ZIP SSS allows improving user options through the following additional characteristics: SPTA SP failure logging; optimization of SP count and accounting for the their characteristic features for the purpose of SP emergency delivery (ED); capability to use products with any type of redundancy; when using SPTA-G group set, capability to include differently-structured products into ESs. The paper sets forth the structure diagram of the DIALOG SSS programs interaction, that implies three modes of operation of the simulation model: SI calculation for specific SPTA contents; calculation of preliminary SPTA supply before the beginning of optimization; calculation of optimal SPTA set. The authors examine the matters related to the selection of the required number and duration of simulation model testing.

Abstract WMP106: Why Are Women Less Represented in ICH Trials?

Background: Fewer women than men tend to be enrolled in clinical trials of intracerebral hemorrhage (ICH). It is unclear whether this reflects lower prevalence of ICH in women, selection bias, poor recruitment efforts, or other factors. We undertook this study to examine differences between women and men in the reasons for exclusion from the iDEF (Intracerebral hemorrhage Deferoxamine) trial. Methods: The screen failure log included 29 different reasons for exclusion. Chi square statistics and p-values were used to evaluate whether women and men differed with regard to reason for screen failure. Findings: The iDEF trial enrolled 294 subjects; 38.5% were women. A total of 8776 subjects were screen failures. Sex was missing in 58. The remaining 8718 were included in this analysis; 3982 women (45.7%) and 4736 men (54.3%) (p&lt;0.0001). The enrollment rates were 2.8% in women vs. 3.7% in men (p=0.01). We were unable to obtain consent in 1.3% of women vs 1.7% of men (p=0.1), and patients/families declined participation in 1.3% of women vs. 1.3% of men (p=0.9). More women than men failed screening because of age &gt;80 (22.4% vs 12.6%) and pre-existing DNR/DNI (3.7% vs. 2.8%). Conversely, fewer women than men failed screening because inability to administer study drug within 24 hour due to late presentation (6.6% vs 7.8%), admission NIHSS score &lt;6 (10.2% vs 13.2%), coagulopathy (5.3% vs 7.5%), inability to comply with the protocol (0.7% vs 1.3%), abnormal renal function (1.9% vs 2.9%), drug/alcohol abuse (1.7% vs 3.7%), and presentation with confirmed aspiration or pneumonia (1.1% vs 1.8%).These differences were statistically significant. Interpretation: Results from this multi-center, prospective, ICH trial indicate that lower rates of women enrollment may be attributed to older age and higher rates of pre-existing DNR/DNI orders. Inability to obtain consent or declining participation was similar between women and men, arguing against selection bias. Our findings should be confirmed in other ICH trials to determine if additional efforts are needed to improve women’s participation in future studies. Funding: US National Institutes of Health and US National Institute of Neurological Disorders and Stroke (U01NS074425)

P4516Incremental prognostic values of cognitive impairment diagnosed by mini-mental state examination and mini-cog in older hospitalized patients with heart failure

Abstract Introduction Cognitive impairment (CI) is associated with worse prognosis in patients with heart failure, especially in the elderly; however, its incremental prognostic ability in pre-existing prognostic models has not been well elucidated. Moreover, although some tools have been proposed for evaluating cognitive function, their difference in prognostic prediction has not been explicitly compared. Methods A total of 352 heart failure patients aged ≥75 years admitted to three hospitals were evaluated for their cognitive function using the Mini-Mental State Examination (MMSE) and Mini-cog during index hospitalization. We diagnosed CI if MMSE and Mini-cog were ≤23 and ≤2, respectively. The primary endpoint was all-cause death. Results The median age of the entire cohort was 85 (IQR: 80–88) years, and 47.7% of the subjects were male. Based on the MMSE and Mini-cog, the CI was diagnosed in 167 (47.4%) and 159 (45.2%) patients, respectively. The two diagnostic tools showed poor to moderate agreement (Cohen's kappa coefficient: 0.37, 95% CI: 0.27–0.47). During the follow-up period of median 346 (IQR: 195–489) days, 53 patients (15.1%) died. Although the Kaplan-Meier analysis showed that CI diagnosed using Mini-cog (CI-MC) was associated with significantly higher mortality (P=0.001), this association was not significant for CI diagnosed using MMSE (CI-MMSE) (P=0.059). On multivariate Cox regression analysis, CI-MMSE and CI-MC were individually associated with worse prognosis in older heart failure patients even after adjustment for Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) risk model and log B-type natriuretic peptide levels (CI-MMSE, HR: 2.05 [95% CI: 1.16–3.61]; and CI-MC, HR: 2.57 [95% CI: 1.46–4.53]). The receiver operating characteristic curve analysis for Mini-cog showed significantly higher area under the curve (AUC) than that for MMSE (0.61 vs. 0.52, p=0.045). To test the incremental prognostic capability, models were constructed by individually adding each score to the MAGGIC risk model, and the net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were evaluated. CI-MMSE did not show incremental prognostic predictability (NRI: 0.28, p=0.069; IDI: 0.01, p=0.090), whereas CI-MC (NRI: 0.45, p=0.001; IDI: 0.03, p=0.001) did. Adding CI-MC instead of CI-MMSE to the MAGGIC risk model showed significant reclassification improvement (NRI: 0.45, p=0.002, IDI: 0.02, p=0.041). Conclusion In older patients with heart failure, CI defined by Mini-Cog is superior in providing additive prognostic value than that defined by CI based on MMSE. Acknowledgement/Funding This study is partially funded by Japan Heart Foundation Research Grant and Novartis Research Grants.

Improving Scalability and Reward of Utility-Driven Self-Healing for Large Dynamic Architectures

Self-adaptation can be realized in various ways. Rule-based approaches prescribe the adaptation to be executed if the system or environment satisfies certain conditions. They result in scalable solutions but often with merely satisfying adaptation decisions. In contrast, utility-driven approaches determine optimal decisions by using an often costly optimization, which typically does not scale for large problems. We propose a rule-based and utility-driven adaptation scheme that achieves the benefits of both directions such that the adaptation decisions are optimal, whereas the computation scales by avoiding an expensive optimization. We use this adaptation scheme for architecture-based self-healing of large software systems. For this purpose, we define the utility for large dynamic architectures of such systems based on patterns that define issues the self-healing must address. Moreover, we use pattern-based adaptation rules to resolve these issues. Using a pattern-based scheme to define the utility and adaptation rules allows us to compute the impact of each rule application on the overall utility and to realize an incremental and efficient utility-driven self-healing. In addition to formally analyzing the computational effort and optimality of the proposed scheme, we thoroughly demonstrate its scalability and optimality in terms of reward in comparative experiments with a static rule-based approach as a baseline and a utility-driven approach using a constraint solver. These experiments are based on different failure profiles derived from real-world failure logs. We also investigate the impact of different failure profile characteristics on the scalability and reward to evaluate the robustness of the different approaches.

Towards comprehensive dependability-driven resource use and message log-analysis for HPC systems diagnosis

Failure analysis plays an important role in the reliability of data centers and high-performance computing (HPC) systems. Recent work have shown that both resource use data and failure logs can, separately and together, be used to detect system failure-inducing errors and diagnose system failures; the result of error propagation and (unsuccessful) execution of error recovery mechanisms. For more accurate and detailed failure diagnosis, knowledge of error propagation patterns and unsuccessful error recovery is important. To improve system reliability, knowledge of recovery protocols deployment is important. This paper describes and demonstrates application of a new diagnostics framework (CORRMEXT). CORRMEXT analyzes and reports error propagation patterns and degrees of success and failure of error recovery protocols. The steps in the framework are correlations of resource use metrics and error messages, and identification of the earliest times of change of system behaviour. The framework is illustrated with analyses of resource use data and message logs for three HPC systems operated by the Texas Advanced Computing Center (TACC). The illustrations are focused on groups of resource use counters and groups of errors; they reveal many interesting insights into patterns of: (i) network data and software errors, (ii) Lustre file-system and Linux operating system process errors, and (iii) memory and storage errors. We also confirm that: (i) correlations of resource use and errors can only be identified by applying different correlation algorithms, and (ii) the earliest times of change in system behaviour can only be identified by analyzing both the correlated resource use counters and correlated errors. We believe CORRMEXT is the first tool that have diagnosed error propagation paths and error recovery attempts on three different HPC systems. CORRMEXT will be put on the public domain to support systems administrators in diagnosing HPC system failures, on August 2018.

Quality Analysis of Water Network Failure

Abstract The main objective of the study was to analyse and assess the failure rate of the water supply system of a provincial city located in south-eastern Poland. In the analysis the exploitation data provided by the water supply company were used. The received data include, among others, material structure, age of water supply network and failure log. The failure rate analysis of the water supply network was based on the failure rate index λ [failure/(km∙year)]. Based on the performed analysis, it was found that the water company should consider renovating or replacing steel pipes.

Understanding Practical Tradeoffs in HPC Checkpoint-Scheduling Policies

As the scale of High-Performance Computing (HPC) clusters continues to grow, their increasing failure rates and energy consumption levels are emerging as serious design concerns. Efficiently running systems at such large scales critically relies on deploying effective, practical methods for fault tolerance while having a good understanding of their respective performance and energy overheads. The most commonly used fault tolerance method is checkpoint/restart. Checkpoint scheduling policies, however, have been traditionally optimized and analysed from one angle: application performance. In this work, we provide an extensive analysis of the performance, energy and I/O costs associated with a wide array of checkpointing policies. We consider practical deployment issues and show that simple formulas can be used to accurately estimate wasted work in a system. We propose methods to optimize checkpoint scheduling for energy savings and evaluate the runtime-optimized and energy-optimized policies using simulations based on failure logs from 10 production HPC clusters. Our results show ample room for achieving high quality energy/performance tradeoffs when using methods that exploit characteristics of real world failures. We also analyze the impact of energy-optimized checkpointing on the storage subsystem and identify policies that are optimal for I/O savings.

Failure Prognosis of High Voltage Circuit Breakers with Temporal Latent Dirichlet Allocation

The continual accumulation of power grid failure logs provides a valuable but rarely used source for data mining. Sequential analysis, aiming at exploiting the temporal evolution and exploring the future trend in power grid failures, is an increasingly promising alternative for predictive scheduling and decision-making. In this paper, a temporal Latent Dirichlet Allocation (TLDA) framework is proposed to proactively reduce the cardinality of the event categories and estimate the future failure distributions by automatically uncovering the hidden patterns. The aim was to model the failure sequence as a mixture of several failure patterns, each of which was characterized by an infinite mixture of failures with certain probabilities. This state space dependency was captured by a hierarchical Bayesian framework. The model was temporally extended by establishing the long-term dependency with new co-occurrence patterns. Evaluation of the high voltage circuit breakers (HVCBs) demonstrated that the TLDA model had higher fidelities of 51.13%, 73.86%, and 92.93% in the Top-1, Top-5, and Top-10 failure prediction tasks over the baselines, respectively. In addition to the quantitative results, we showed that the TLDA can be successfully used for extracting the time-varying failure patterns and capture the failure association with a cluster coalition method.

Screen failure data in clinical trials: Are screening logs worth it?

Clinical trials frequently spend considerable effort to collect data on patients who were assessed for eligibility but not enrolled. The Consolidated Standards of Reporting Trials (CONSORT) guidelines' recommended flow diagram for randomized clinical trials reinforces the belief that the collection of screening data is a necessary and worthwhile endeavor. The rationale for collecting screening data includes scientific, trial management, and ethno-socio-cultural reasons. We posit that the cost of collecting screening data is not justified, in part due to inability to centrally monitor and verify the screening data in the same manner as other clinical trial data. To illustrate the effort and site-to-site variability, we analyzed the screening data from a multicenter, randomized clinical trial of patients with transient ischemic attack or minor ischemic stroke (Platelet-Oriented Inhibition in New Transient Ischemic Attack and Minor Ischemic Stroke (POINT)). Data were collected on over 27,000 patients screened across 172 enrolling sites, 95% of whom were not enrolled. Although the rate of return of screen failure logs was high overall (95%), there were a considerable number of logs that were returned with 'no data to report' (23%), often due to administrative reasons rather than no patients screened. In spite of attempts to standardize the collection of screening data, due to differences in site processes, multicenter clinical trials face challenges in collecting those data completely and uniformly. The efforts required to centrally collect high-quality data on an extensive number of screened patients may outweigh the scientific value of the data. Moreover, the lack of a standardized definition of 'screened' and the challenges of collecting meaningful characteristics for patients who have not signed consent limits the ability to compare across studies and to assess generalizability and selection bias as intended.

Board-Level Functional Fault Diagnosis Using Artificial Neural Networks, Support-Vector Machines, and Weighted-Majority Voting

Increasing integration densities and high operating speeds lead to subtle manifestation of defects at the board level. Functional fault diagnosis is, therefore, necessary for board-level product qualification. However, ambiguous diagnosis results lead to long debug times and even wrong repair actions, which significantly increase repair cost and adversely impact yield. Advanced machine-learning (ML) techniques offer an unprecedented opportunity to increase the accuracy of board-level functional diagnosis and reduce high-volume manufacturing cost through successful repair. We propose a smart diagnosis method based on two ML classification models, namely, artificial neural networks (ANNs) and support-vector machines (SVMs) that can learn from repair history and accurately localize the root cause of a failure. Fine-grained fault syndromes extracted from failure logs and corresponding repair actions are used to train the classification models. We also propose a decision machine based on weighted-majority voting, which combines the benefits of ANNs and SVMs. Three complex boards from the industry, currently in volume production, and additional synthetic data, are used to validate the proposed methods in terms of diagnostic accuracy, resolution, and quantifiable improvement over current diagnostic software.