It is essential to allocate virtual machines efficiently in order to optimize the resource utilization and minimize the total energy consumption in data centers (DCs). Thus, reducing the number of operational physical machines will reduce the overall energy consumption of the cloud DC. In this paper, First Fit (FF), First Fit Decrease (FFD), Best Fit (BF), Best Fit Decrease (BFD) and Genetic Algorithm (GA) heuristics have been used to allocate virtual machines (VMs) to physical machines (PMs) to get the optimal mapping. Furthermore, we estimated the total energy consumption by the resource requirements of PMs that are required to process the assigned VMs. The experimental results indicate that optimizing the number of active PMs is not always enough to minimize the total energy consumption of the DC. GA is able to get the optimal number of active PMs, but it does not always reduce the total energy consumption of the DC compared to the other employed algorithms. This highlights the necessity of considering energy consumption as a separate objective during virtual machine allocation and consolidation planning as opposed to minimizing the number of active PMs to reduce the power consumption of the data center.

Given the rise in demand for cloud computing in the modern era, the effectiveness of resource utilization is eminent to decrease energy footprint and achieve economic services. With the emerging machine learning and artificial intelligence techniques to model and predict, it is essential to explore a principal method that provides the best solution for the accurate provisioning of forthcoming requests in a cloud data center. Recent studies used machine learning and other advanced analytics to predict resource usage; however, these do not consider long-range dependencies in the time series, which is essential to capture for better prediction. Further, they show limitations in handling noise, missing values, and outliers in datasets. In this paper, we explored the problem by studying three techniques that enabled us to answer improvements in short-term forecasting of physical machines’ resource usage if the above factors are considered. We evaluated the predictions using Transformer and Informer deep learning models that cover the above aspects and compared them with the Long short-term memory (LSTM) model. We used a real-world Google cluster trace usage dataset and employed Balanced Iterative Reducing and Clustering using Hierarchies (BIRCH) algorithm to select heterogeneous machines. The evaluation of the three models depicts that the Transformer architecture that considers long-range dependencies in time series and shortcomings with datasets shows improvement in forecasting with 14.2% reduction in RMSE than LSTM. However, LSTM shows better results for some machines than the Transformer, which depicts the importance of input sequence order. The Informer model, which considers both dependencies and is a hybrid of LSTM and Transformer, outperformed both models with 21.7% from LSTM and 20.8% from Transformer reduction in RMSE. The results also depict Informer model consistently performs better than the other models across all subsets of the dataset. Our study proves that considering long-range dependencies and sequence ordering for resource usage time series improves the prediction.

The World Health Organization (WHO) has declared the novel coronavirus as global pandemic on 11 March 2020. It was known to originate from Wuhan, China and its spread is unstoppable due to no proper medication and vaccine. The developed forecasting models predict the number of cases and its fatality rate for coronavirus disease 2019 (COVID-19), which is highly impulsive. This paper provides intrinsic algorithms namely - linear regression and long short-term memory (LSTM) using deep learning for time series-based prediction. It also uses the ReLU activation function and Adam optimiser. This paper also reports a comparative study on existing models for COVID-19 cases from different continents in the world. It also provides an extensive model that shows a brief prediction about the number of cases and time for recovered, active and deaths rate till January 2021.

Nowadays, Cloud services are gaining importance among users due to their cost-effectiveness and highly scalable resources. To meet the user’s demands, several data centres are built across the globe, which has severe environmental as well as economical concerns. Energy consumption is one of the most significant issue faced by cloud service providers. Prediction of accurate resource usage of the physical machine helps in effective utilization of resources in a data centre, resulting in minimizing an active number of physical machines, which helps to minimize the energy consumption of a data centre. Although several models till date focus on virtual machine consolidation with a notion of reducing energy consumption, the reduction of operational physical machines has not gathered enough attention. In this paper, we propose a prediction model to predict resource utilization of physical machines, which enables to effectively utilize the entire data centre’s resources to reduce energy consumption. First, the raw time series workload is processed to enhance the value of its features for better training and prediction of mean resource utilization in the cloud data centre using the proposed Sum Average (SA) algorithm. Afterward, Gaussian Mixture Model (GMM) is employed to cluster heterogeneous machines of data centre based on its resource usage which helps to analyze the prediction for each kind of configured machine available in a data centre. In addition, the Long Short Term Memory model (LSTM) is employed to predict the mean resource usage of physical machines for every clustered machine. Furthermore, the effectiveness of our proposed model is evaluated using the Google cluster trace usage dataset. Lastly, the proposed model is compared with Linear Regression, Moving Average, and Auto Regression Integrated Moving Average model. Root Mean Square Error (RMSE) analysis states that our proposed model outperforms the other compared techniques

The paper considers the problem of an effective microservices interaction and its organization to support data consistency in fault tolerant and high load systems. The “Saga” microservices orchestration template was used for microservices management. The authors assessed the expediency of asynchronous programming principles usage for designing the Saga coordinator. The simulation of processes in the Saga coordinator was conducted; it considers different specifics of asynchronous and synchronous configurations of distributed transactions (Sagas) management. Synchronous configuration group includes a coordinator with a fixed pool of threads and a coordinator that generates a new thread for each new Saga. Asynchronous configuration group consists of a coroutines based coordinator and a coordinator that uses Linux core scheduler. The set of simulations with different numbers of Sagas and available coordinator processors was executed. It was shown that the use of asynchronous approaches significantly reduces the Saga’s execution duration up to 9.74 times and improves the processor time utilisation value up to 88 %. The obtained data proves the efficiency of asynchronous programming principles applied to the design of the Saga coordinator. The difference in efficiency between the asynchronous algorithms that were implemented in this paper was insignificant. Asynchronous programming principles used to build the Saga coordinator allow it to handle bigger load and to use processor resources in a more efficient way. The outcome of this research can be applied during the design of fault tolerant and high load systems. The paper might be interesting to IT-specialists and researchers focusing on distributed computing.

In the popular field of cloud computing, millions of job requests arrive at the data centre for execution. The job of the data centre is to optimally allocate virtual machines (VMs) to these job requests in order to use resources efficiently. In the future smart cities, huge amount of job requests and data will be generated by the Internet of Things (IoT) devices which will influence the designing of optimum resource management of smart cloud environments. The present paper analyses the performance efficiency of the data centre with and without job request consolidation. First, the work load performance of the data centre was analysed without job request consolidation, exhibiting that the job requests to VM assignment was highly imbalanced, and only 5% of VMs were running with a load factor of more than 70%. Then, the technique for order of preference by similarity to ideal solution-based VM selection algorithm was applied, which was able to select the best VM using parameters such as the provisioned or available central processing unit capacity, provisioned or available memory capacity, and state of machine (running, hibernated, or available). The Bitbrains dataset consisting of 1750 VMs was used to analyse the performance of the proposed methodology. The analysis concluded that the proposed methodology was capable of serving all job requests using less than 24% VMs with improved load efficiency. The fewer number of VMs with an improved load factor guarantees energy saving and an increase in the overall running efficiency of the smart data centre environment.

Conventional machine learning (ML) needs centralized training data to be present on a given machine or datacenter. The healthcare, finance, and other institutions where data sharing is prohibited require an approach for training ML models in secured architecture. Recently, techniques such as federated learning (FL), MIT Media Lab's Split Neural networks, blockchain, aim to address privacy and regulation of data. However, there are difference between the design principles of FL and the requirements of Institutions like healthcare, finance, etc., which needs blockchain-orchestrated FL having the following features: clients with their local data can define access policies to their data and define how updated weights are to be encrypted between the workers and the aggregator using blockchain technology and also prepares audit trail logs undertaken within network and it keeps actual list of participants hidden. This is expected to remove barriers in a range of sectors including healthcare, finance, security, logistics, governance, operations, and manufacturing.

Microservices RESTful architecture is almost a standard for e-commerce web applications today. It brings domain isolation, development and support independence. In the same time it increases complexity of cross-domain interactions. The case when distributed changes should not break consistency of microservice states is one of challenging task that might appear during development of such systems. Defining sagas in central orchestrator that performs changes in microservices one by one and controls compensations in case if failure occurs is one of the well-known approaches today. It solves the problem of consistency but creates new vulnerable area in fault tolerant environment. Usage of saga cluster and additional optimizations of its structure are modeled, evaluated and proposed in this paper. Provided fault tolerant solution with improved time and memory characteristics.

Nowadays computer modeling area is very popular and mostly researches interest in cloud computing in computer modeling rapidly growing. In scientific world math models with high complexity are continuously developed in different areas of applications, that is a cause of this growth. In this way our team should quick respond for new user requirements and growing of computation complexity for system dynamics models. Previously software architecture of sdCloud platform was complex and hard to maintain and extend. The new architecture based on micro-services infrastructure and Enterprise Service bus provide flexibility, scalability and reliability of computation platform. In this paper we describe all details of software architecture that was we built for sdCloud platform to compute system dynamics models. The paper paid special attention to the communication process of services with each other using the Enterprise Service Bus and introduce a new term – services responsibilities zone.

This article describes an approach for designing a storage system that will allow persistence of large number of heterogeneous entities, used to describe software development process in various process supporting tools (bug trackers, repositories, wiki engines, etc). Goal of this work is to extract core semantics essences from entities like issues, code commits, builds in continuous integration systems and others and persist them together in one indexable, integrated and searchable systems disregarding from which tool each of them came from. Taking into high level of variety in data shapes and forms, requirement to perform complicated cross-entities queries and potentially high volumes of data -designing a storage for such systems turns into a challenging task.

Nowadays, different kinds of modelling settled down in most areas of human activities. Often, it is not easy to design and create a model of an existing systems, because equations linking its components are not known. This article describes an approach allowing to identify system components interaction based on run-time monitoring and analysis of system behavior. Described method is shown on a simple system that will be turned in to a system dynamics mode by automated equations definition and results will compared to initial analytic model for the same system.

This paper presents a process of adaptation of system dynamics models execution algorithms to cloud-based environment. System dynamics is an aspect of systems theory as a method to understand the dynamic behaviour of complex systems. Existing modeling algorithms used in popular modeling solutions are either not available for free use or have several disadvantages which prevent them from being used in distributed cloud environment. Adaptation of execution algorithms aimed not only to adapt execution process to distributed parallel environments with higher reliability and wider range of possible applications, but also to improve system dynamics model execution performance. For example, existing algorithms of model execution which are not ready for distributed environments will fail to complete modeling task in case of hardware failure, and optimized ones are able to smoothly transfer execution process from one node to another with minimal impact on overall model execution progress. Such capabilities help to save many resources and, especially, time on execution re-runs. In this paper described algorithms and approaches designed for sdCloud solution which are focused on transferring execution of system dynamics models into distributed cloud-based environment and shown extra benefits brought to modeling process by shift to the cloud.

Nowadays, the amount of both the data themselves and their sources is constantly growing. However, such diversity often leads to heterogeneity and fragmentation of data, which makes complete analysis difficult. At the same time, the multimodal data approach is able to address these issues and not only to find new links in different data sets, but also to calculate the coherence power of their proximity. The approach described in the article is extremely important for the organisation and preliminary analysis of data, but it also can be used as an independent method for various data analysis activities.

This article addresses estimation of source code changes influence on a high-scale software solutions. Described existing approaches address this issue only partially while the proposed solution is aiming to provide an ability of deep influence tracing from changed source code piece through the rest of the project. Such information can be used as a solid and reliable baseline for the risk management and estimation in cases of making code changes in the project or when changing version of used third party library with available source codes.

Nowadays System Dynamics is one of the most popular modelling approaches because it provides instruments for simulation and understanding of complicated processes in a wide ranch of application domains. In consequence of this, modellers, who are working in this area, are dealing with big number of entities both, common to the modelling approach and specific for their particular application area. By its nature, system dynamics approaches allow to investigate overall behaviour of the system or process. In the same time many of relations between particular model nodes remains unclear. Described multimodal approach allows to dive deeper in to the insights of each particular model participant and helps to attempt to reveal inner effects make to each other.

This article describes a new approach for system dynamics models execution. In most cases when model execution is involved it is performed on a set of static and known data, which are sent to the model as an input. And it is expected, that on the model output modeler will get a set of other system or event characteristics, computed by the model based on the input parameters. This approach still has the widest usage, but it is not the only one scenario, which is demanded by different industries. With growing popularity of concepts such as Internet of Things, demand in modeling based solutions, which take as input continuous data streams, has grown significantly. In comparison with stand-alone client-side modeling systems, cloud-based solutions, such as sdCloud, became a reasonable answer to such industry request. Such systems can provide an ability of continuous execution of system dynamics models. In other words, these systems are ready to accept an incoming data stream and perform model execution that will result in streaming modeling results back to the end-user. Running system dynamics models in parallel with the process it is describing allows to perform predictive modeling of the system status in the future, and it also allows to find additional hidden external impacts to the model. For example, such approach can be a base for predictive maintenance of complicated technical systems, because it allows computing nearest maintenance time more efficient.

This chapter dedicated to analysis of various types of data produced or processed by Internet of Things solution with ESB architecture accent. Since Internet of Things platforms are mostly focused on gathering and processing big amounts of data, to keep such solutions efficient it is important to design and create efficient storage mechanism that will not became a bottle-neck for the whole system performance. Taking into account types of data to be stored or processed, in the chapter designed an option for building efficient storage mechanism for Internet of Things solution that flexibly fits various data requirements.

This article describes a new project - sdCloud, which was designed in purpose to create an effective cloud-based execution environment for System Dynamics models. The article contains overview of basic terms and conditions in System Dynamics, description of sdCloud project and investigations of issues, which has occurred in development process, including integration of existing System Dynamics tools, possibilities of optimization processes in System Dynamics modeling and its basic approaches.

This paper describes a new open source project (sdCloud) dedicated to creating a cloud-based execution environment for system dynamics models. The goal is to provide model sharing and remote model execution and result generation. This solution is targeted to become publicly available at the end of June 2016.

System Dynamics ReviewVolume 31, Issue 1-2 p. 86-95 Notes and Insights SDXchange: stand-alone translators to enable XMILE model adaptation, transportation, and exchange Robert Ward, Corresponding Author Robert Ward Iowa State UniversityCorrespondence to: Robert Ward, Iowa State University E-mail: [email protected]Search for more papers by this authorJames Houghton, James Houghton MIT Sloan School of ManagementSearch for more papers by this authorIvan A. Perl, Ivan A. Perl National Research University of Information Technologies, Mechanics and Optics (ITMO), Saint Petersburg, RussiaSearch for more papers by this author Robert Ward, Corresponding Author Robert Ward Iowa State UniversityCorrespondence to: Robert Ward, Iowa State University E-mail: [email protected]Search for more papers by this authorJames Houghton, James Houghton MIT Sloan School of ManagementSearch for more papers by this authorIvan A. Perl, Ivan A. Perl National Research University of Information Technologies, Mechanics and Optics (ITMO), Saint Petersburg, RussiaSearch for more papers by this author First published: 09 September 2015 https://doi.org/10.1002/sdr.1529Read the full textAboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Volume31, Issue1-2January-June 2015Pages 86-95 RelatedInformation