Search

Traditional cloud architectures struggle to effectively allocate resources to container-based workloads due to fluctuating usage patterns and potential interference among multi-tenants. Conventional scheduling methods, which primarily rely on user-specified resource requests, often lead to over-provisioning and suboptimal resource utilization. Although efforts have been made to predict container resource usage and allocate resources more tightly than the full requests, such approaches typically fall short during sudden demand spikes, thus failing to meet Service Level Objectives (SLOs). In this article, we introduce a novel cloud resource prediction engine specifically designed to differentiate between online and batch jobs. Our engine prioritizes ensuring SLOs for online jobs where immediate responsiveness is crucial. Specifically, our approach employs a combination of genetic algorithms (GA) and Bayesian neural networks (BNN) to enhance the prediction accuracy of CPU and memory resources. Trained on real-world trace data, our model significantly outperforms traditional forecasting methods like ARIMA and exponential smoothing, especially in reducing the risk of under-prediction for online jobs. This not only ensures more efficient resource utilization but also improves adherence to SLOs without compromising performance. [ABSTRACT FROM AUTHOR]

Nowadays, with the development of science and technology, as well as the increase of important data and transactions, maintaining these data and transactions has become a big challenge. On the one hand, their maintenance cost is a very important issue for organizations and companies, and on the other hand, their security and safety is a very important and sensitive issue because the occurrence of software faults, especially Byzantine faults, hardware faults and cyber-attacks, threaten data and transactions and the safety of systems. Therefore, researchers are trying to provide solutions that can provide the best service at the lowest cost according to the pay-as-you-go law and can maintain the security and health of data in the event of a fault. One of the most important techniques presented to increase fault tolerance in distributed systems is the use of replication methods, which besides being costly, have many problems. In this article, blockchain technology is used to achieve goals such as increasing reliability and availability, reducing resources, reducing costs, and increasing fault tolerance, especially Byzantine faults, and has achieved very good results compared to other methods. [ABSTRACT FROM AUTHOR]

Container virtualization technology represented by Docker has been widely used in the industry due to its advantages of lightweight, fast deployment, and easy portability. This paper considers the scenarios of AI-based IoT applications based on container technology in a cloud-edge collaborative environment, and proposes a container-based task scheduling algorithm. Using priority-aware greedy strategy, a new scheduling algorithm named PGT has been proposed which adopts the multi-criteria approach Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). The containers in cloud servers and edge servers are managed in a uniform platform, and IoT application services are deployed in containers. The task with smaller deadline constraint is scheduled first due to its higher priority. Then, multiple indicators are considered comprehensively, such as task response time, energy consumption, task execution cost, to find the optimal container to execute task. Through varying the number of edge servers and the number of tasks, the simulation results in a cloud-edge collaborative environment indicate that the proposed scheduling approach outperforms the four baseline algorithms in improving QoS satisfaction rate, energy consumption, penalty cost and total violation time. [ABSTRACT FROM AUTHOR]

Containerization technology utilizes operating system level virtualization to package applications to run with required libraries and are isolated from other processes on the same host. Lightweight and quick deployment make containers popular in many data centers. Running distributed applications in data centers usually involves multiple clusters of machines. Docker Swarm is a container orchestration tool for managing a cluster of Docker containers and their hosts. However, Docker Swarm's scheduler does not consider resource utilization when placing containers in a cluster. This paper first investigated performance interference in container clusters. Our experimental study showed that distributed applications' performance can be degraded when co-located with other containers which aggressively consume resources. A new scheduler is proposed to improve performance while keeping high resource utilization. The experimental results demonstrated that the proposed prototype with machine learning based clustering algorithms could effectively improve distributed applications' performance by up to 14.5% with an average at around 12%. This work also provides theoretical bounds for the container placement problem. [ABSTRACT FROM AUTHOR]

Cloud computing, with elasticity and pay-as-you-go pricing, is a suitable platform for executing workflow applications. Workflow as a Service (WaaS) systems provide scientists with an easy-to-use, and cost-effective platform to execute their workflow applications in the cloud at any time or location worldwide. Quality of Service (QoS) is recognized as a key requirement in WaaS. Monetary cost and time are two primary QoS from a clients' perspective; whereas, energy consumption is considered a significant problem for cloud providers' profitability and ability to provide low-cost services. Most workflow scheduling studies assume that workflow tasks have a deterministic Execution Time (ET), which is generally unrealistic in the real world. However, there are few approaches for scheduling in WaaS considering deadlines, and monetary costs with uncertain task ET. These studies typically assume that a cloud resource can execute all types of workflow applications without any need for additional software components. However, using containers is a suitable solution to provide an executable environment for the execution of any workflow type on cloud resources. To this end, we present two cost and energy-aware workflow scheduling that consider the uncertainty in tasks' ETs. Both solutions are designed for WaaS, leveraging containers to enhance resource utilization rate and reduce energy consumption, resource monetary cost, and workflows deadline violations. Simulated experiments demonstrate that our proposed methods outperform two recent state-of-the-art scheduling algorithms in terms of success rate, monetary cost, energy consumption, and resource utilization rate. [ABSTRACT FROM AUTHOR]

There are many organizations interested in sharing data with others, and they can do this only if a multi-domain secure platform is available. Such platforms, often referred to as Digital Data Marketplaces (DDMs), require that all the transactions follow the pre-defined policies that are established by the participating parties i.e, domains. However, building a multi-domain network infrastructure in which each domain can manage its own connectivity while at the same time all of the transactions follow the sharing agreements is still a challenge. In this paper, we introduce a multi-domain containerized DDM that is built upon a P4-based network. It can handle the communication of multiple domains and guarantee that the operation of transactions is based on the pre-defined policies. We also studied the setup performance by defining a model which we demonstrated follows the real measurements, and we can use for decision making. The results also show the low overhead of using P4 switch in network setup time. In addition, we conducted a security evaluation which showed that our P4-based network setup is secure against most types of attacks. [ABSTRACT FROM AUTHOR]

The Compact Muon Solenoid (CMS) experiment heavily relies on the CMSWEB cluster to host critical services for its operational needs. The cluster is deployed on virtual machines (VMs) from the CERN OpenStack cloud and is manually maintained by operators and developers. The release cycle is composed of several steps, from building RPMs to their deployment, validation, and integration tests. To enhance the sustainability of the CMSWEB cluster, CMS decided to migrate its cluster to a containerized solution based on Docker and orchestrated with Kubernetes (K8s). This allows us to significantly speed up the release upgrade cycle, follow the end-to-end deployment procedure, and reduce operational cost. In this paper, we give an overview of the CMSWEB VM cluster and the issues we discovered during this migration. We discuss the architecture and the implementation strategy in the CMSWEB Kubernetes cluster. Even though Kubernetes provides horizontal pod autoscaling based on CPUs and memory, in this paper, we provide details of horizontal pod autoscaling based on the custom metrics of CMSWEB services. We also discuss automated deployment procedure based on the best practices of continuous integration/continuous deployment (CI/CD) workflows. We present performance analysis between Kubernetes and VM based CMSWEB deployments. Finally, we describe various issues found during the implementation in Kubernetes and report on lessons learned during the migration process. [ABSTRACT FROM AUTHOR]

Container-based virtualization is a new technology used by cloud providers to provide cloud services to end-user. This technology has various advantages (e.g. lightweight, quickly deployable, and efficient for resource utilization) for executing an application. It reduces the operating cost, carbon emission, and allocates the resources dynamically. Different cloud applications have different requirements. Deploying resources according to peak requirements always can be costly. On the other hand, always having minimum computing resources may not meet workload's peak requirements, and may cause degraded system performance, less throughput, more response time and service level agreement violations. Hence, it becomes a challenge to maintain optimal level of resources to fulfill the SLA requirements for the applications. To address the above issues, we propose an auto-scaler which uses proactive approach (Support Vector Regression) to perform horizontal elasticity for Docker containers in response to fluctuating workload for real-time applications. As the workload increases, additional resources will be allocated dynamically supporting elasticity. The increase in capacity of a machine dynamically is termed as elasticity. The effective mechanism of elasticity avoids the violation of SLA and penalties in terms of user's loss. The proposed auto-scaler uses the IBM computing model, MAPE-K principle to perform elasticity using the workload predictions made by the SVR model. The predicted workload helps auto-scaler to find out the minimum numbers of replicas needed for a container of a cluster so that it handles the future workload. The experimental results show that the results of SVM prediction keep the performance of the system sustainable with fluctuating workload. [ABSTRACT FROM AUTHOR]

Autoscaling of containers can leverage performance measures from the different layers of the computational stack. This paper investigate the problem of selecting the most appropriate performance measure to activate auto-scaling actions aiming at guaranteeing QoS constraints. First, the correlation between absolute and relative usage measures and how a resource allocation decision can be influenced by them is analyzed in different workload scenarios. Absolute and relative measures could assume quite different values. The former account for the actual utilization of resources in the host system, while the latter account for the share that each container has of the resources used. Then, the performance of a variant of Kubernetes' auto-scaling algorithm, that transparently uses the absolute usage measures to scale-in/out containers, is evaluated through a wide set of experiments. Finally, a detailed analysis of the state-of-the-art is presented. [ABSTRACT FROM AUTHOR]

Virtualization is a foundational element of cloud computing. Since cloud computing is slower than a native system, this study analyzes ways to improve performance. We compared the performance of Docker and Kernel-based virtual machine (KVM). KVM uses full virtualization, including × 86 hardware virtualization extensions. Docker is a solution provided by isolation in userspace instead of creating a virtual machine. The performance of KVM and Docker was compared in three ways. These comparisons show that Docker is faster than KVM. [ABSTRACT FROM AUTHOR]

The setup environment and deployment of distributed applications is a human intensive and highly complex process that poses significant challenges. Nowadays many applications are developed in the cloud and existing applications are migrated to the cloud because of the promising advantages of cloud computing. Presenting two common serious challenging scenarios in the application development environment, we propose a multi-task PaaS cloud infrastructure using Docker and AWS services for application isolation, optimization and rapid deployment of distributed applications. We fully utilized Docker, a lightweight containerization technology that uses a host of the Linux kernel's features such as namespaces and cgroup's to sandbox processes into configurable virtual environments. The Amazon EC2 container service helps our container management framework. The cluster management framework uses optimistic, shared state scheduling to execute processes on EC2 instances using Docker containers. Several experimentations were carried out, one of the experimentation focused on a simulation of application deployment scheduling that shows our propose infrastructure is flexible, efficient and well optimized. [ABSTRACT FROM AUTHOR]