Cloud computing is a widely adopted platform for executing tasks of different application types that belong to the end users. In the cloud, application task is prone to failure for several reasons, such as software bug or exception, virtual or physical infrastructure failure. Cloud service providers are responsible for managing availability of scheduled computing tasks in order to provide high level QoS for their customers. Protecting task against failure is a challenging and not a trivial mission due to dynamic, heterogeneous and large distributed structure of the cloud environment. The existing works in the literature focus on task failure prediction and neglect the remedy (post) actions. In this work, we first study and analyze three publicly available large cluster datasets from Google, Alibaba, and Trinity, to characterize task failure in cloud computing platform. We then propose a failure-aware task scheduling framework that can predict the termination status for a set of given tasks during the runtime, and take the appropriate remedy actions. The framework uses deep learning methods named Artificial and Convolutional Neural Network, ANN and CNN, for different prediction purposes. In addition, we formalize the actions selection problem as Integer Linear Programming (ILP) model and propose a heuristic optimization solution that aims to minimize the failure probability of tasks and their resources usage. The results show ANN and CNN can achieve prediction accuracy of up to 94% and 92%, respectively using Google dataset. Moreover, the framework can protect up to 40% of tasks that are predicted as failed using Alibaba dataset by taking the appropriate remedy actions, and hence save many of cluster’s resources such as CPU and RAM.

Containers are recognized for their lightweight and virtualization efficiency, making them a vital element in modern application orchestration. In this context, the scheduler is crucial in strategically distributing containers across diverse computing nodes. This paper presents a novel two-stage container scheduling solution that addresses node imbalances and efficiently deploys containers. The proposed solution formulates the scheduling process as an optimization problem, integrating various objective functions and constraints to enhance server consolidation and minimize energy consumption. The confidentiality of migrated containers is ensured through encryption, and the associated costs are incorporated into the optimization constraints. This approach ensures security in container scheduling, considering container attributes as input features in our proposed attributes-based encryption model. By carefully selecting containers and destination nodes, this work seeks to establish balance within cloud-based clusters. This contributes to the improvement of container orchestration systems and their effectiveness in real-world scenarios. The proposed solution’s efficacy is demonstrated in its ability to efficiently deploy containers in multi-data center cloud environments and seamlessly migrate them between hosts within the same data center or across different data centers. Our results show optimal consolidation with a reduction in the number of running hosts, ranging from 4% to over 18%. Additionally, the solution promotes minimal total power consumption with savings ranging from 3.5 to 16.25 megawatts, while also ensuring balanced server loads.

Cloud elasticity involves timely provisioning and de-provisioning of computing resources and adjusting resources size to meet the dynamic workload demand. This requires fast, and accurate resource scaling methods at minimum cost (e.g. pay as you go) that match with workload demands. Two dynamic changing parameters must be defined in an elastic model, the workload resource demand classes, and the data center resource reconfiguration classes. These parameters are not labeled for cloud management system while data center logs are being captured. Building an advance elastic model is a critical task, which defines multiple classes under these two categories i.e. for workload and for provisioning. A dynamic method is therefore required to define (during configuration time window) the workload classes and resource provisioning classes. Unsupervised learning model such as K-Means has many challenges such as time complexity, selection of optimum number of clusters (representing the classes), and determining centroid values of the clusters. All clustering methods depend on minimizing mean square error between center of population in same class member. These methods are often enhanced using guidelines to find out the centroids, but they suffer from K-Means limitations. For the application of clustering cloud log traces, most of the reported work use K-Means clustering to label workload types. However, there is no work reported that label data center scaling classes. In this work, a novel method is proposed to analyze the characteristics of both workloads and datacenter configurations using clustering method, and is based on random variable model transformation (kernel density estimator) guide. This method enhances K-Means clustering by automatically determining optimum number of classes and finding the mean centroids for the clusters. In addition, it improves the accuracy and the time complexity of standard K-Means clustering model, by best correlating between clustering attributes using statistical correlation methods.

Predicting workload demands can help to achieve elastic scaling by optimizing data center configuration, such that increasing/decreasing data center resources provides an accurate and efficient configuration. Predicting workload and optimizing data center resource configuration are two challenging tasks. In this work, we investigate workload and data center modeling to help in predicting workload and data center operation that is used as an experimental environment to evaluate optimized elastic scaling for real data center traces. Three methods of machine learning are used and compared with an analytical approach to model the workload and data center actions. Our approach is to use an analytical model as a predictor to evaluate and test the optimization solution set and find the best configuration and scaling actions before applying it to the real data center. The results show that machine learning with an analytical approach can help to find the best prediction values of workload demands and evaluate the scaling and resource capacity required to be provisioned. Machine learning is used to find the optimal configuration and to solve the elasticity scaling boundary values. Machine learning helps in optimization by reducing elastic scaling violation and configuration time and by categorizing resource configuration with respect to scaling capacity values. The results show that the configuration cost and time are minimized by the best provisioning actions.