Workload Adaptation Research Articles

A particle swarm optimization inspired global and local stability driven predictive load balancing strategy

In distributed systems and parallel computing, optimal load balancing is difficult. These abstract addresses load balancing in distributed situations, highlighting current solutions' flaws and emphasizing the need for new ones. Load balancing research includes centralized and distributed algorithms, heuristics, and predictive models. Despite various successful methods, workload adaptability, overhead reduction, and scaling to large systems remain unresolved. This study proposes a particle swarm optimization (PSO) load balancing method that considers global and local stability considerations. The proposed method uses PSO principles to balance exploration and exploitation and allocate resources among distributed nodes. Predictive components improve preventative load management by predicting workload changes. Global and local load balancing stability criteria distinguish this study. The recommended method considers global system-wide performance indicators, local node-level characteristics, and micro-level stability to maximize system efficiency. A dual-focus technique distinguishes the proposed load balancing strategy from others, solving dynamic distributed system challenges. The study examines load balancing system advances and suggests improvements and further research. More accurate prediction modeling, stability measures, and application-specific enhancements may be studied in the future. Experimental validation and real-world implementation of the recommended approach are necessary to determine its practicality and ability to handle modern distributed computing systems.

SplitStream: Distributed and workload-adaptive video analytics at the edge

Deep learning-based video analytics is computation-intensive. Manufacturers such as Nvidia have launched many embedded deep learning accelerators and are rapidly gaining market share. However, the computing resources of such accelerators are still limited and heterogeneous. Although existing systems aim at optimizing video query tasks from a variety of perspectives, they rarely consider the general cooperation between heterogeneous edge devices and the dynamic workload of video content. In this work, we present SplitStream, a distributed system for accelerating video query tasks across heterogeneous edge devices, which is able to fully utilize the resources on each device and adapt to the workload dynamics. The key to achieving this is the data parallelism brought by the multi-instance mechanism and the dynamic workload adaptability brought by the two-level workload balancing mechanism. Evaluation results show SplitStream reduces the result retrieval time by up to 19% and improves the resource utilization by up to 234%.

ERGONOMICS AND OCCUPATIONAL PHYSIOLOGY: A MULTIPROFESSIONAL APPROACH TO WORK

In ergonomics, approaches, perspectives and purposes in relation to work are presented from a multiprofessional nature, insofar as it covers disciplines that bring contributions to ergonomics in its application in the reality of work, with improvement of operating conditions, relocation, reform and application of ergonomic projects. The research problem presents the following question: What are the multidisciplinary approaches used in the context of ergonomics in companies to reduce physiological and biomechanical damage in the work environment? The objective of this article is to identify in ergonomics and occupational physiology the approaches, perspectives and purposes in relation to the work of a multiprofessional nature. It was intended to demonstrate the concept of ergonomics in work relationships; indicate the interdisciplinary and multiprofessional nature of ergonomics; to point out the biomechanical and physiological aspects in the context of organizational ergonomics. The assumptions of the study indicate that the application of ergonomics and its approaches is necessary in organizations in the creation of projects in technologicalenvironments where repetitive activities generate occupational diseases, requiring prevention mechanisms through the creation of ergonomic projects aimed at the nature of work in situations that require corrective action. From the use of methods and steps, ergonomics allows a new perception of the functioning of a company, from the understanding of the work carried out to conceive the work situations in which workers have control over physical, mental and psychic balance and for an improvement of the system. In order to respond to the objectives of the study and the problem pointed out, it was decided to carry out a bibliographic research based on the theoretical foundations of the authors. The results showed that ergonomics deals with the transformation of the work environment for the biomechanical and physiological adaptation to the type of work activity. Therefore, it must be part of the company's culture, favoring transformations from the insertion of ergonomic projects to the optimization of human work, quality and productivity in a safe and healthy environment. Regarding the biomechanical and physiological aspects of organizational ergonomics, there is a need to evaluate the postural dynamics of the human body and its direct relationship with the workload and efficient adaptation to avoid diseases.

Adaptive Workload Allocation for Multi-Human Multi-Robot Teams for Independent and Homogeneous Tasks

Multi-human multi-robot (MH-MR) systems have the ability to combine the potential advantages of robotic systems with those of having humans in the loop. Robotic systems contribute precision performance and long operation on repetitive tasks without tiring, while humans in the loop improve situational awareness and enhance decision-making abilities. A system's ability to adapt allocated workload to changing conditions and the performance of each individual (human and robot) during the mission is vital to maintaining overall system performance. Previous works from literature including market-based and optimization approaches have attempted to address the task/workload allocation problem with focus on maximizing the system output without regarding individual agent conditions, lacking in real-time processing and have mostly focused exclusively on multi-robot systems. Given the variety of possible combination of teams (autonomous robots and human-operated robots: any number of human operators operating any number of robots at a time) and the operational scale of MH-MR systems, development of a generalized framework of workload allocation has been a particularly challenging task. In this paper, we present such a framework for independent homogeneous missions, capable of adaptively allocating the system workload in relation to health conditions and work performances of human-operated and autonomous robots in real-time. The framework consists of removable modular function blocks ensuring its applicability to different MH-MR scenarios. A new workload transition function block ensures smooth transition without the workload change having adverse effects on individual agents. The effectiveness and scalability of the system's workload adaptability is validated by experiments applying the proposed framework in a MH-MR patrolling scenario with changing human and robot condition, and failing robots.

A Deep Learning Method for Bearing Fault Diagnosis through Stacked Residual Dilated Convolutions

Real-time monitoring and fault diagnosis of bearings are of great significance to improve production safety, prevent major accidents, and reduce production costs. However, there are three primary concerns in the current research, namely real-time performance, effectiveness, and generalization performance. In this paper, a deep learning method based on stacked residual dilated convolutional neural network (SRDCNN) is proposed for real-time bearing fault diagnosis, which is subtly combined by the dilated convolution, the input gate structure of long short-term memory network (LSTM) and the residual network. In the SRDCNN model, the dilated convolution is used to exponentially increase the receptive field of convolution kernel and extract features from the sample with more points, alleviating the influence of randomness. The input gate structure of LSTM could effectively remove noise and control the entry of information contained in the input sample. Meanwhile, the residual network is introduced to overcome the problem of vanishing gradients caused by the deeper structure of the neural network, hence improving the overall classification accuracy. The experimental results indicate that compared with three excellent models, the proposed SRDCNN model has higher denoising ability and better workload adaptability.

Low-Power Scalable 3-D Face Frontalization Processor for CNN-Based Face Recognition in Mobile Devices

A low-power scalable 3-D face frontalization processor is proposed for accurate face recognition in mobile devices. In spite of recent improvement in face recognition accuracy mainly from convolutional neural networks (CNNs), their performance is limited to face images with frontal view. For face recognition with human-level accuracy in real-life environment, in which most of the face images are captured from arbitrary angles, 3-D face frontalization is essential as a preprocessing stage for CNN-based face recognition algorithms. The proposed face frontalization processor shows scalability in two aspects: image resolution and accuracy. For low-power consumption and scalability, the processor proposes three features: 1) scalable processing element (PE) architecture with workload adaptation; 2) accuracy scalable regression weight quantization to reduce the external memory access (EMA) down to 81.3%; and 3) pipelined memory-level zero-skipping to further reduce the EMA by 98.4% without any latency overhead. From the proposed EMA reduction features, the EMA is reduced by 99.7% with little accuracy degradation in face frontalization results. The proposed face frontalization processor is implemented in 65-nm CMOS process, and it shows 4.73 frames/s) throughput. Moreover, power consumption of the implemented face frontalization processor is 0.53 mW, which is suitable for applications on mobile devices.

Autonomic workload performance tuning in large-scale data repositories

The workload in large-scale data repositories involves concurrent users and contains homogenous and heterogeneous data. The large volume of data, dynamic behavior and versatility of large-scale data repositories is not easy to be managed by humans. This requires computational power for managing the load of current servers. Autonomic technology can support predicting the workload type; decision support system or online transaction processing can help servers to autonomously adapt to the workloads. The intelligent system could be designed by knowing the type of workload in advance and predict the performance of workload that could autonomically adapt the changing behavior of workload. Workload management involves effectively monitoring and controlling the workflow of queries in large-scale data repositories. This work presents a taxonomy through systematic analysis of workload management in large-scale data repositories with respect to autonomic computing (AC) including database management systems and data warehouses. The state-of-the-art practices in large-scale data repositories are reviewed with respect to AC for characterization, performance prediction and adaptation of workload. Current issues are highlighted at the end with future directions.

A survey and experimental comparison of distributed SPARQL engines for very large RDF data

Distributed SPARQL engines promise to support very large RDF datasets by utilizing shared-nothing computer clusters. Some are based on distributed frameworks such as MapReduce; others implement proprietary distributed processing; and some rely on expensive preprocessing for data partitioning. These systems exhibit a variety of trade-offs that are not well-understood, due to the lack of any comprehensive quantitative and qualitative evaluation. In this paper, we present a survey of 22 state-of-the-art systems that cover the entire spectrum of distributed RDF data processing and categorize them by several characteristics. Then, we select 12 representative systems and perform extensive experimental evaluation with respect to preprocessing cost, query performance, scalability and workload adaptability, using a variety of synthetic and real large datasets with up to 4.3 billion triples. Our results provide valuable insights for practitioners to understand the trade-offs for their usage scenarios. Finally, we publish online our evaluation framework, including all datasets and workloads, for researchers to compare their novel systems against the existing ones.

A Model of Security Adaptation for Limited Resources in Wireless Sensor Network

View of wireless sensor network (WSN) devices is small but have exceptional functionality. Each node of a WSN must have the ability to compute and process data and to transmit and receive data. However, WSN nodes have limited resources in terms of battery capacity, CPU, memory, bandwidth, and data security. Memory limitations mean that WSN devices cannot store a lot of information, while CPU limitations make them operate slowly and limited battery capacity makes them operate for shorter periods of time. Moreover, the data gathered and processed by the network face real security threats. This article presents an Adaptable Resource and Security Framework (ARSy) that is able to adapt to the workload, security requirements, and available resources in a wireless sensor network. The workload adaptation is intended to preserve the resource availability of the WSN, while the security adaptation balances the level of security with the resource utilization. This solution makes resources available on the basis of the workload of the system and adjusts the level of security for resource savings and makes the WSN devices work more efficiently.

On-demand re-optimization of integration flows

Integration flows are used to propagate data between heterogeneous operational systems or to consolidate data into data warehouse infrastructures. In order to meet the increasing need of up-to-date information, many messages are exchanged over time. The efficiency of those integration flows is therefore crucial to handle the high load of messages and to reduce message latency. State-of-the-art strategies to address this performance bottleneck are based on incremental statistic maintenance and periodic cost-based re-optimization. This also achieves adaptation to unknown statistics and changing workload characteristics, which is important since integration flows are deployed for long time horizons. However, the major drawbacks of periodic re-optimization are many unnecessary re-optimization steps and missed optimization opportunities due to adaptation delays. In this paper, we therefore propose the novel concept of on-demand re-optimization. We exploit optimality conditions from the optimizer in order to (1) monitor optimality of the current plan, and (2) trigger directed re-optimization only if necessary. Furthermore, we introduce the PlanOptimalityTree as a compact representation of optimality conditions that enables efficient monitoring and exploitation of these conditions. As a result and in contrast to existing work, re-optimization is immediately triggered but only if a new plan is certain to be found. Our experiments show that we achieve near-optimal re-optimization overhead and fast workload adaptation.

QoC elastic scheduling for real-time control systems

The elastic task model, a significant development in scheduling of real-time control tasks, provides a mechanism for flexible workload management in uncertain environments. It tells how to adjust the control periods to fulfill the workload constraints. However, it is not directly linked to the quality-of-control (QoC) management, the ultimate goal of a control system. As a result, it does not tell how to make the best use of the system resources to maximize the QoC improvement. To fill in this gap, a new feedback scheduling framework, which we refer to as QoC elastic scheduling, is developed in this paper for real-time process control systems. It addresses the QoC directly through embedding both the QoC management and workload adaptation into a constrained optimization problem. The resulting solution for period adjustment is in a closed-form expressed in QoC measurements, enabling closed-loop feedback of the QoC to the task scheduler. Whenever the QoC elastic scheduler is activated, it improves the QoC the most while still meeting the system constraints. Examples are given to demonstrate the effectiveness of the QoC elastic scheduling.

I/O scheduling and performance analysis on multi‐core platforms

AbstractMulti‐core platforms are being widely used in many technology areas. These technology shifts not only provide a compute‐intensive service, but also require us to improve the I/O‐intensive service in performance. However, the existing I/O scheduling approaches lack either quantitative analysis or workload adaptability. These approaches become worse when the application scenarios have either diverse I/O tasks or heterogeneous multi‐cores. Based on the analysis of traditional software pipelining technology, we first propose a reverse interleaved pipelining scheduling strategy to decrease the total I/O execution time and balance the workloads of homogeneous processors in this paper. Theoretical analysis shows that this balancing strategy is fair. As for the heterogeneous processors' circumstance, a Self‐adaptive scheduling method is presented for selecting one of the most‐appropriate multi‐core services for the expected I/O tasks. Simulation studies show that the performance of this scheduling strategy self‐adapts to application and multi‐core diversity well. Copyright © 2009 John Wiley & Sons, Ltd.

Evaluating Dynamic Task Mapping in Network Processor Runtime Systems

Modern network processor systems require the ability to adapt their processing capabilities at runtime to changes in network traffic. Traditionally, network processor applications have been optimized for a single static workload scenario, but recently several approaches for runtime adaptation have been proposed. Comparing these approaches and developing novel runtime support algorithms is difficult due to the multicore system-on-a-chip nature of network processors. In this paper, we present a model for network processors that can aid in evaluating different runtime support systems. The model considers workload characteristics of applications and network traffic using a queuing network abstraction. The accuracy of this analytical approach to modeling runtime systems is validated through simulation. We illustrate the effectiveness of our model by comparing the performance of two existing workload adaptation algorithms.

The interesting cross-paths of HIV/AIDS and water in Southern Africa with special reference to South Africa

HIV/AIDS and water-borne diseases account for a substantial degree of morbidity and mortality in different age groups across the globe, but their ripple effects are more devastating in developing countries. Estimates of the HIV/AIDS epidemic in South Africa vary but attest to a mature and generalised epidemic. Antenatal sero-prevalence data show a rising epidemic from less than 1% prevalence among pregnant women attending antenatal clinics in 1990 to over 29.5% in 2004. Prevalence is generally higher among females than males, indicating feminisation of the epidemic. Race analyses in 2003 reveal the highest prevalence to be among African (13%) followed by Whites and Coloureds (6.2%) and least among Indians (1.6%). The epidemic which has left a wave of orphans, child-headed households and its attendant ripple effects on household structures, demography, economy and education is characterised by a ‘pigeon-hole\' of several opportunistic infections. In rural areas, devoid of electricity and potable water, the impact is more profound because of the role of water in cooking, drinking, consumption of anti-retrovirals and in the preparation of milk supplements for infants. People with compromised immune systems are more prone to several diseases than individuals whose immune systems are not compromised by HIV/AIDS. HIV/AIDS patients therefore have greater requirements for potable water than uninfected individuals. Improving water quality will lead to a decline in child and adult mortality as well as diarrhoeal diseases in people living with HIV/AIDS. One of the hallmarks of HIV/AIDS is diarrhoea and about 90% of HIV/AIDS patients in Africa suffer from chronic diarrhoea, an easily recognisable clinical manifestation of water-borne infections. Other linkages between HIV/AIDS and water are breast feeding; unsafe water used in infant formula preparation which increases the risk of diarrhoeal diseases and deaths in infants; anti-retroviral treatment, as safe water is needed during anti-retroviral treatment or treatment for opportunistic infections in HIV. In addition to providing safe water, supply points and latrines should be close to points of use as this will reduce the long distances that care givers and HIV/AIDS patients undertake to fetch water and appropriate water system design is required. Added to the above-mentioned are poverty alleviation aspects; poverty enhances vulnerability to HIV infection and AIDS exacerbates poverty. Secured access to sufficient safe water and sanitation is also a human rights issue, just as stigmatisation of HIV/AIDS has made the disease a human rights issue. The cross-paths between HIV/AIDS and water have long-term implications for effective water resource management and the provision of wholesome water to communities. Such implications include faltering payment for water supply because HIV and AIDS are financially dis-empowering, erosion of social capital and waning productivity. Mainstreaming of HIV/AIDS in the water sector is of utmost importance, including development of work-place policies, adaptation and reorganisation of workload, development of strategies for reserve staff, adjustment of performance appraisal systems, pro-poor financing with a focus on water for health and economic benefit and integration of HIV and AIDS into training activities. Water SA Vol.32 (3) 2006: pp.323-343

Enabling structural summaries for efficient update and workload adaptation

To facilitate queries over semi-structured data, various structural summaries have been proposed. Structural summaries are derived directly from data and serve as the indexes for evaluating path expressions. We introduce D( k)-index, an adaptive structural summary, for general graph-structured data. Building on previous 1-index and A( k)-index, D( k)-index is also based on the concept of bisimilarity. However, as a generalization of 1-index and A( k)-index, D( k)-index possesses the adaptive ability to adjust its structure to changes in query load. It also enables efficient update algorithms, which are crucial to real applications but have not been adequately addressed in previous literatures. Our experiments show that D( k)-index is a more effective structural summary than previous static ones as a result of its query load sensitivity. In addition, the update operations on it can be performed more efficient than on its predecessors.

Time and space adaptation for computational grids with the ATOP-Grid middleware

Adaptive resource allocation is becoming an important feature for local applications and for grid applications that run simultaneously on multiple sites. The applications may face varying resource availability during execution, may need to schedule around obstacles as from reservation, and may have to deal with varying system load under time-shared execution and with a lack of accurate runtime predictability on heterogeneous resources. Thus, middleware support is needed to make the applications adaptable. We present our ATOP-Grid middleware which employs the well-known Zoltan/ParMETIS library for high-quality workload adaptation, enhanced with an over-partitioning extension. ATOP-Grid constitutes a unified approach for application-internal workload self-adaptation under different resource-sharing types, integration with the local batch job schedulers, and largely autonomous decision making.

Workload and Strategic Adaptation under Transformations of Visual-Coordinative Mappings

Six right-handed male volunteers performed a bi-manual pursuit tracking task. Unknown to the subject, the mapping between both the absolute and relative motions of each control lever changed between trials, and could also change within a trial. There were four conditions which remained constant within a trial and four others which switched between combinations of the first four mappings. These eight total mappings could be performed under either a fast or slow target speed giving sixteen possible conditions, the order of administration of which was randomized across the subjects. Performance was assessed through the use of root mean square error (RMS) and percent time-on-target (TOT). Workload was derived from subjective report using the NASA Task-Load Index. Performance results differentiated three different groups within the mappings. The first group represented conditions in which despite scaling factors, the movement of the cursor remained consistent with the movement of the controls and was directly mapped. The second group, consisting of only two conditions, also presented internal consistency except that the movements of the cursor were mirrored in the movements of the controls, a direct but inverted map. The third group represented those conditions in which at least one period of performance decoupled the actions of the two controls. In general, workload distinguished the ordering of difficulty of the conditions, except that there was evidence of dissociation between performance and subjective evaluation in the three most difficult decoupled conditions. Verbal report of performer strategic behavior was elicited. Each line of evidence is considered in the attempt to understand how subjects explored the perceptual-motor space for strategic solutions when presented with the autonomy to do so.