Input Workload Research Articles

Closed-loop workload input–output control of production systems: A hybrid simulation study

Workload Control (WLC) is a method of adjusting the overall workload in the production system by controlling the input of production orders and the output of produced items. It contributes to the predictability of lead times and more accurate delivery date commitments in make-to-order manufacturing. Input control relates to order release and output control to capacity adjustment. In a recent previous work, a novel approach to integrated input–output control has been proposed: a dynamic closed-loop model where automatic feedback control is added and where order release and capacity decisions are based on the observed shop floor state. In this paper, this approach is further investigated in a more realistic hybrid simulation setting, combining discrete-time controllers and a discrete-event model of a manufacturing shop. The hybrid model was applied to two different systems: an elementary 4-machine shop floor (as a proof of concept) and a more complex real system, a manufacturing cell of cylinder liners. The results show that simultaneous input–output control is effective and enables the systems to maintain WIP balance, absorb demand fluctuations, and effectively reject disturbances. The simulations of the complex system showed that it might be advantageous to provide more local or more global information to the controllers, depending on the variability of the processing times and on the managerial focus. Closed-loop discrete-event models for WLC production control are unprecedented. Also, a simultaneous input and output control setting is much less reported in WLC literature than settings with input control only (i.e., order release only). This paper contributes in both ways.

StreamOps: Cloud-Native Runtime Management for Streaming Services in ByteDance

Stream processing is widely used for real-time data processing and decision-making, leading to tens of thousands of streaming jobs deployed in ByteDance cloud. Since those streaming jobs usually run for several days or longer and the input workloads vary over time, they usually face diverse runtime issues such as processing lag and varying failures. This requires runtime management to resolve such runtime issues automatically. However, designing a runtime management service on the ByteDance scale is challenging. In particular, the service has to concurrently manage cluster-wide streaming jobs in a scalable and extensible manner. Furthermore, it should also be able to manage diverse streaming jobs effectively. To this end, we propose StreamOps to enable cloud-native runtime management for streaming jobs in ByteDance. StreamOps has three main designs to address the challenges. 1) To allow for scalability, StreamOps is running as a standalone lightweight control plane to manage cluster-wide streaming jobs. 2) To enable extensible runtime management, StreamOps abstracts control policies to identify and resolve runtime issues. New control policies can be implemented with a detect-diagnose-resolve programming paradigm. Each control policy is also configurable for different streaming jobs according to the performance requirements. 3) To mitigate processing lag and handling failures effectively, StreamOps features three control policies, i.e., auto-scaler, straggler detector, and job doctor, that are inspired by state-of-the-art research and production experiences at ByteDance. In this paper, we introduce the design decisions we made and the experiences we learned from building StreamOps. We evaluate StreamOps in our production environment, and the experiment results have further validated our system design.

COMPREHENSIVE METHOD OF ENERGY-EFFICIENT WORKLOAD PROCESSING IN THE INFORMATION AND COMMUNICATION NETWORK

Background. Peculiarities of the workload in a modern information and communication network (ICN) determine specific requirements for energy efficiency, performance and availability of its processing system. Existing approaches to increase energy efficiency and performance of workload processing do not take into account the possibility of dynamic changes in ICN workload arrival rate and individual energy consumption characteristics of computing nodes of the system.
 Objective. The purpose of the paper is to increase the energy efficiency and performance of ICN workload processing while meeting the requirements for the availability of the processing system, taking into account dynamic changes in the input workload arrival rate and the individual characteristics of the computing nodes’ energy consumption.
 Methods. A mathematical model of the workload processing system was built using the queueing theory methods, and an ontological model of this system was built using intelligent data analysis methods, which made it possible to quantitatively and qualitatively describe the complex relationships between system parameters and workload processing efficiency indicators. On the basis of the built models, a comprehensive method of energy-efficient workload processing was proposed, which differs from the known ones by the use of individual energy consumption models of computing nodes, combining the advantages of horizontal scaling and energy-efficient scheduling taking into account dynamic changes in workload arrival rate.
 Results. The efficiency of ICN workload processing is increased by 15.722% according to the efficiency criterion, which includes energy efficiency and performance indicators, compared to the known energy-efficient Backfill approach while meeting the requirements for the availability of the processing system.
 Conclusions. The energy efficiency, performance and availability of the ICN workload processing system can be improved by combining horizontal scaling and energy-efficient scheduling approaches using individual energy consumption models of computing nodes and taking into account dynamic changes in the input workload arrival rate.

Theoretical Analysis of an Adaptive Periodic Multi Installment Scheduling With Result Retrieval for SAR Image Processing

Processing a large-scale Synthetic Aperture Radar (SAR) image dataset on a distributed computing infrastructure poses a challenging problem. Large-scale load distribution strategies like multi-installment scheduling (MIS) assume that the size of the result is negligible compared to the input workloads and hence ignore it in their design. Similarly, numerical methods like particle swarm optimization and their variants are not practical for real-time applications, given their run-time complexities. As both the results retrieval and completion time are crucial for SAR image data processing, in this article, we attempt to provide a thorough theoretical analysis of an adaptive MIS that includes the result retrieval phase. We use the periodic nature of the internal installments to keep the strategy simple and fine-tune the last installment to avoid any idle times in the processors. We derive a closed-form solution for the load fractions and hence, the overall processing time, schedule feasibility criteria, and certain other properties that lead to adaptive scheduling. Finally, we validate our theoretical findings through rigorous simulation studies using a loosely connected virtual machines (VMs) topology for the SAR dataset.

Tunable Encrypted Deduplication with Attack-resilient Key Management

Conventional encrypted deduplication approaches retain the deduplication capability on duplicate chunks after encryption by always deriving the key for encryption/decryption from the chunk content, but such a deterministic nature causes information leakage due to frequency analysis. We present TED , a tunable encrypted deduplication primitive that provides a tunable mechanism for balancing the tradeoff between storage efficiency and data confidentiality. The core idea of TED is that its key derivation is based on not only the chunk content but also the number of duplicate chunk copies, such that duplicate chunks are encrypted by distinct keys in a controlled manner. In particular, TED allows users to configure a storage blowup factor, under which the information leakage quantified by an information-theoretic measure is minimized for any input workload. In addition, we extend TED with a distributed key management architecture and propose two attack-resilient key generation schemes that trade between performance and fault tolerance. We implement an encrypted deduplication prototype TEDStore to realize TED in networked environments. Evaluation on real-world file system snapshots shows that TED effectively balances the tradeoff between storage efficiency and data confidentiality, with small performance overhead.

Mathematical modeling of the workload processing in the information and communication network

The functioning of a modern information and communication network is impossible without software, which partially replaces specialized equipment when solving network tasks. These tasks form a workload that needs to be processed by a distributed server infrastructure within the network. There are requirements for each type of workload regarding the quality of its processing (maximum service time, permissible losses, etc.). Thus, there is a need to establish relationships between quality indicators and parameters of the processing system to ensure the fulfillment of these requirements using mathematical modeling. The paper proposes a mathematical model of a distributed workload processing system in the information and communication network as a queuing system. Unlike known models, the proposed model considers the possible variable nature of the input workload arrival rate and parallelization methods that may be applied to the software that implements it. Within the modeling process, a method of transition from a non-stationary non-ordinary incoming queries’ flow to an ordinary stationary flow is proposed. Based on the constructed mathematical model, a complex method of energy-efficient workload processing has been developed. A laboratory experiment has proven the efficiency of the proposed complex method and the mathematical model’s adequacy in its basis.

DISTILL

Many database systems offer index tuning tools that help automatically select appropriate indexes for improving the performance of an input workload. Index tuning is a resource-intensive and time-consuming task requiring expensive optimizer calls for estimating the cost of queries over potential index configurations. In this work, we develop low-overhead techniques that can be leveraged by index tuning tools for reducing a large number of optimizer calls without making changes to the tuning algorithm or to the query optimizer. First, index tuning tools use rule-based techniques to generate a large number of syntactically-relevant indexes; however, a large proportion of such indexes are spurious and do not lead to a significant improvement in the performance of queries. We eliminate such indexes much earlier in the search by leveraging patterns in the workload, without making optimizer calls. Second, we learn cost models that exploit the similarity between query and index configuration pairs in the workload to efficiently estimate the cost of queries over a large number of index configurations using fewer optimizer calls. We perform an extensive evaluation over both real-world and synthetic benchmarks, and show that given the same set of input queries, indexes, and the search algorithm for exploration, our proposed techniques can lead to a median reduction in tuning time of 3X and a maximum of 12X compared to state-of-the-art tuning tools with similar quality of recommended indexes.

HealthMon: An approach for monitoring machines degradation using time-series decomposition, clustering, and metaheuristics

Monitoring the degradation of machines to anticipate potential failures represents a significant challenge. In Industry 4.0, this task is critical when the costs associated with maintenance and stoppages on the productive processes are high. Nowadays, many preventive maintenance techniques employ supervised or unsupervised machine learning algorithms. However, the definition of which features should be processed by such algorithms is not a simple task, being crucial to the proposed technique’s success. Against this background, we consider whether unsupervised algorithms combined with time-series decomposition can enhance the estimate of a machine’s health. This article proposes HealthMon as a novel approach to compute a health index of machines based on sensor measurements. HealthMon extracts time-series from such sensors, which are decomposed in an unsupervised way to present the health state along time. The health index is related to the degradation of the considered machine, thus optimizing the machine maintenance schedule. This work advances the state-of-the-art in the following ways: (i) it proposes a novel index of machines health, which yields a more direct and intuitive view of machine degradation; (ii) it devises the first approach capable of estimating the health index of a machine in a completely unsupervised way; (iii) it generalizes vibrating and rotating machines, thus being able to monitor a wide range of industrial equipment. We evaluated our method using both simulated and real data. The results show that the evolution of vibrating machines’ failures can be effectively detected under various input workloads. Finally, through HealthMon, industry decision-makers benefit from the guidelines for preventive actions at appropriate times, thus meeting Industry 4.0.

Cosine

We present a self-designing key-value storage engine, Cosine, which can always take the shape of the close to "perfect" engine architecture given an input workload, a cloud budget, a target performance, and required cloud SLAs. By identifying and formalizing the first principles of storage engine layouts and core key-value algorithms, Cosine constructs a massive design space comprising of sextillion (10 36 ) possible storage engine designs over a diverse space of hardware and cloud pricing policies for three cloud providers - AWS, GCP, and Azure. Cosine spans across diverse designs such as Log-Structured Merge-trees, B-trees, Log-Structured Hash-tables, in-memory accelerators for filters and indexes as well as trillions of hybrid designs that do not appear in the literature or industry but emerge as valid combinations of the above. Cosine includes a unified distribution-aware I/O model and a learned concurrency-aware CPU model that with high accuracy can calculate the performance and cloud cost of any possible design on any workload and virtual machines. Cosine can then search through that space in a matter of seconds to find the best design and materializes the actual code of the resulting storage engine design using a templated Rust implementation. We demonstrate that on average Cosine outperforms state-of-the-art storage engines such as write-optimized RocksDB, read-optimized WiredTiger, and very write-optimized FASTER by 53x, 25x, and 20x, respectively, for diverse workloads, data sizes, and cloud budgets across all YCSB core workloads and many variants.

On proposing a non-intrusive device and methodology to monitor motor degradation

Monitoring the degradation of induction motors is an essential concern for industries, especially when there are high costs related to maintenance and their stops. Nowadays, there are many non-invasive ways to monitor the health status of induction motors, including sensors to measure temperature, magnetic field, vibration, current, and power. However, installing these sensors may not be a simple task since intrusive procedures such as engine disassembling, wire separation, and engine interruption are commonly observed. Thus, we propose MUCLA (Multiphase Clamp and Tracking - patent pending) as a new device and methodology for monitoring motor degradation by analyzing its stator current. MUCLA introduces a non-invasive current sensors array on a proposed clamp in the hardware perspective, while a spectral tracking methodology refers to our novelty in the software viewpoint. For end-users, the main MUCLA contribution relies on the combination of the simplicity of installation and the possibility of using the proposed device even when considering Variable-Frequency Drive powered engines. The results obtained with both simulation and practical tests show encouraging results over different input workloads and emulated degradation. Compared to the current initiatives, we can classify MUCLA as a better solution for industrial plants in terms of installation time, installation complexity, and cost.

A Novel Improved Grey Wolf Optimization Algorithm Based Resource Management Strategy for Big Data Systems

Presently, big data is very popular, since it finds helpful in diverse domains like social media, E-commerce transactions, etc. Cloud computing offers services on demand, broader networking access, source collection, quick flexibility and calculated services. The cloud sources are usually different and the application necessities of the end user are rapidly changing from time to time. So, the resource management is the tedious process. At the same time, resource management and scheduling plays a vital part in cloud computing (CC) results, particularly while the environment is employed in the analysis of big data, and minimum predictable workload dynamically enters into the cloud. The identification of the optimal scheduling solutions with diverse variables in varying platform still remains a crucial problem. Under cloud platform, the scheduling techniques should be able to adapt the changes quickly and according to the input workload. In this paper, an improved grey wolf optimization (IGWO) algorithm with oppositional learning principle has been important to carry out the scheduling task in an effective way. The presented IGWO based scheduling algorithm achieves optimal cloud resource usage and offers effective solution over the compared methods in a significant way.

Decentralized Edge-to-Cloud Load Balancing: Service Placement for the Internet of Things

The Internet of Things (IoT) requires a new processing paradigm that inherits the scalability of the cloud while minimizing network latency using resources closer to the network edge. Building up such flexibility within the edge-to-cloud continuum consisting of a distributed networked ecosystem of heterogeneous computing resources is challenging. Furthermore, IoT traffic dynamics and the rising demand for low-latency services foster the need for minimizing the response time and balanced service placement. Load-balancing for fog computing becomes a cornerstone for cost-effective system management and operations. This paper studies two optimization objectives and formulates a decentralized load-balancing problem for IoT service placement: (global) IoT workload balance and (local) quality of service (QoS), in terms of minimizing the cost of deadline violation, service deployment, and unhosted services. The proposed solution, EPOS Fog, introduces a decentralized multi-agent system for collective learning that utilizes edge-to-cloud nodes to jointly balance the input workload across the network and minimize the costs involved in service execution. The agents locally generate possible assignments of requests to resources and then cooperatively select an assignment such that their combination maximizes edge utilization while minimizes service execution cost. Extensive experimental evaluation with realistic Google cluster workloads on various networks demonstrates the superior performance of EPOS Fog in terms of workload balance and QoS, compared to approaches such as First Fit and exclusively Cloud-based. The results confirm that EPOS Fog reduces service execution delay up to 25% and the load-balance of network nodes up to 90%. The findings also demonstrate how distributed computational resources on the edge can be utilized more cost-effectively by harvesting collective intelligence.

Optimized Sampling Strategies to Model the Performance of Virtualized Network Functions

Modern network services make increasing use of virtualized compute and network resources. This is enabled by the growing availability of softwarized network functions, which take on major roles in the total traffic flow (such as caching, routing or as firewall). To ensure reliable operation of its services, the service provider needs a good understanding of the performance of the deployed softwarized network functions. Ideally, the service performance should be predictable, given a certain input workload and a set of allocated (virtualized) resources (such as vCPUs and bandwidth). This helps to estimate more accurately how much resources are needed to operate the service within its performance specifications. To predict its performance, the network function should be profiled in the whole range of possible input workloads and resource configurations. However, this input can span a large space of multiple parameters and many combinations to test, resulting in an expensive and overextended measurement period. To mitigate this, we present a profiling framework and a sampling heuristic to help select both workload and resource configurations to test. Additionally, we compare several machine-learning based methods for the best prediction accuracy, in combination with the sampling heuristic. As a result, we obtain a reduced dataset which can still model the performance of the network functions with adequate accuracy, while requiring less profiling time. Compared to uniform sampling, our tests show that the heuristic achieves the same modeling accuracy with up to five times less samples.

Latency‐aware adaptive micro‐batching techniques for streamed data compression on graphics processing units

SummaryStream processing is a parallel paradigm used in many application domains. With the advance of graphics processing units (GPUs), their usage in stream processing applications has increased as well. The efficient utilization of GPU accelerators in streaming scenarios requires to batch input elements in microbatches, whose computation is offloaded on the GPU leveraging data parallelism within the same batch of data. Since data elements are continuously received based on the input speed, the bigger the microbatch size the higher the latency to completely buffer it and to start the processing on the device. Unfortunately, stream processing applications often have strict latency requirements that need to find the best size of the microbatches and to adapt it dynamically based on the workload conditions as well as according to the characteristics of the underlying device and network. In this work, we aim at implementing latency‐aware adaptive microbatching techniques and algorithms for streaming compression applications targeting GPUs. The evaluation is conducted using the Lempel‐Ziv‐Storer‐Szymanski compression application considering different input workloads. As a general result of our work, we noticed that algorithms with elastic adaptation factors respond better for stable workloads, while algorithms with narrower targets respond better for highly unbalanced workloads.

Toward Identifying Consumer Responsibility on Voltage Distortion and Filtering Based on the Maclaurin Series

Reducing the voltage distortion is an essential concern for consumers and power distribution companies. The distortion voltage due to the distorted current of consumers affects the common coupling point and consequently, other consumers. Nowadays, most of the harmonic-based active filtering initiatives are generic and act to minimize the harmonics independently of their origin, either the distribution companies and consumers. In this context, we propose SRAF (Shared Responsibility in Active Filtering) as a new approach to determine the portion of responsibility for the voltage distortions, which also implies on defining a novel technique to mitigate the disturbance caused only by the consumer. The main contribution of SRAF relies on the combination of the simplicity of the method in terms of physics understanding and the use of MacLaurin series with on-the-fly parameter optimization. Thus, we introduce in the literature a new easy-to-use form to understand energy responsibility on voltage distortion. The results obtained by simulation show encouraging results, even with different input workloads and simulated environments. By using the SRAF approach massively, we expect lower financial costs in filtering systems. Since our approach observes the consumer viewpoint, we emphasize that contingency filtering is also essential in the case of distortions from the power grid.

Formalizing and simulating cross-layer elasticity strategies in Cloud systems

Clouds are complex systems that provide computing resources in an elastic way. Elasticity allows their adaptation to input workloads by (de)provisioning resources as the demand rises and drops. Given the numerous overlapping factors that impact their elasticity and the unpredictable nature of the workload, providing accurate action plans to manage Cloud elasticity is a particularly challenging task. In this paper, we propose a formal approach based on bigraphical reactive systems to model Cloud structures and their elastic behavior. We design cross-layer elasticity strategies which operate at application and infrastructure Cloud layers to manage the elastic adaptations. We encode the elastic behaviors in Rewriting logic, through the Maude framework, to enable their generic executability. We provide a qualitative verification of the designed behaviors’ correctness with a model-checking technique supported by the linear temporal logic. Finally, we provide a tooled simulation-based methodology, through the Queuing theory, to conduct a quantitative analysis of the designed elasticity strategies.

Workload and Capacity Optimization for Cloud-Edge Computing Systems with Vertical and Horizontal Offloading

A collaborative integration between cloud and edge computing is proposed to be able to exploit the advantages of both technologies. However, most of the existing studies have only considered two-tier cloud-edge computing systems which merely support vertical offloading between local edge nodes and remote cloud servers. This paper thus proposes a generic architecture of cloud-edge computing with the aim of providing both vertical and horizontal offloading between service nodes. To investigate the effectiveness of the design for different operational scenarios, we formulate it as a workload and capacity optimization problem with the objective of minimizing the system computation and communication costs. Because such a mixed-integer nonlinear programming (MINLP) problem is NP-hard, we further develop an approximation algorithm which applies a branch-and-bound method to obtain optimal solutions iteratively. Experimental results show that such a cloud-edge computing architecture can significantly reduce total system costs by about 34%, compared to traditional designs which only support vertical offloading. Our results also indicate that, to accommodate the same number of input workloads, a heterogeneous service allocation scenario requires about a 23% higher system costs than a homogeneous scenario.

Electroencephalographic Workload Indicators During Teleoperation of an Unmanned Aerial Vehicle Shepherding a Swarm of Unmanned Ground Vehicles in Contested Environments.

Background: Although many electroencephalographic (EEG) indicators have been proposed in the literature, it is unclear which of the power bands and various indices are best as indicators of mental workload. Spectral powers (Theta, Alpha, and Beta) and ratios (Beta/(Alpha + Theta), Theta/Alpha, Theta/Beta) were identified in the literature as prominent indicators of cognitive workload.Objective: The aim of the present study is to identify a set of EEG indicators that can be used for the objective assessment of cognitive workload in a multitasking setting and as a foundational step toward a human-autonomy augmented cognition system.Methods: The participants' perceived workload was modulated during a teleoperation task involving an unmanned aerial vehicle (UAV) shepherding a swarm of unmanned ground vehicles (UGVs). Three sources of data were recorded from sixteen participants (n = 16): heart rate (HR), EEG, and subjective indicators of the perceived workload using the Air Traffic Workload Input Technique (ATWIT).Results: The HR data predicted the scores from ATWIT. Nineteen common EEG features offered a discriminatory power of the four workload setups with high classification accuracy (82.23%), exhibiting a higher sensitivity than ATWIT and HR.Conclusion: The identified set of features represents EEG indicators for the objective assessment of cognitive workload across subjects. These common indicators could be used for augmented intelligence in human-autonomy teaming scenarios, and form the basis for our work on designing a closed-loop augmented cognition system for human-swarm teaming.

Pods – A novel intelligent energy efficient and dynamic frequency scalings for multi-core embedded architectures in an IoT environment

In the Advent of the Internet of Things (IoT), embedded architecture takes an important dimension in terms of energy and accomplishment. The embedded system needs more and more intelligent algorithms for better performance and energy efficiency to fit into an IoT scenario. Moreover, with the existence of high-performance multi-core embedded architectures, achievements of energy efficiency remains in the dark side of the research. Several algorithms such as dynamic frequency scaling, thread mapping, starvation methodologies were proposed in embedded architectures for efficient usages of clock frequencies and these features were used as the energy saving modes in which the consumption of energy in the embedded architectures are being controlled. But these methods have several backlogs which permits the use of consumption in the embedded architectures. Considering the above features, this paper proposes a new methodology PODS(Predictors for Optimized Dynamic Scaling) which integrates a powerful machine learning algorithm for scaling the clock frequencies by the input workloads and allocation of the core depending based on the workload. The proposed framework PODS has different phases of working namely workload extraction, characterization, and optimization using BAT algorithms and prediction extreme Machine - Learning. The algorithm was tested on ARM/Cortex architectures (Raspberry Pi 3 Model B+), an evaluated algorithm using the IoMT benchmarks and various parameters that include energy consumption, accuracy of detection/prediction was determined and analyzed. It is found that the implementation of the proposed framework in the test is seen resulting between 35 and 40% reduction in the consumption of the power.

Women's empowerment: the case of smallholder rice farmers in Kilombero District, Tanzania

ABSTRACTThis study examines women’s empowerment and its determinants for smallholder rice farming households in Kilombero, Tanzania. The Women Empowerment in Agriculture Index (WEAI) is adopted and for the study site, the overall WEAI was 0.54 with a 5 Domains Empowerment sub-index value of 0.50 and the Gender Parity sub-index of 0.86. Key domains contributing to women’s disempowerment are workload, resource ownership and restricted inputs to productive decision making. To assess determinants of women’s empowerment the ordinal logit analysis is used and for the female-headed households, age of the household head, education level, group membership, condition of dwelling and distance from the nearest major town have a positive association with women’s empowerment, while monthly income has a negative association with it. For the male-headed households, the association is significant for age of the man.