In the past few years, many researchers attempted to tackle the problem of decreasing energy consumption in cloud data centers. One of the widely adopted techniques for this purpose is dynamic Virtual Machine (VM) consolidation. Consolidation moves VMs between hosts to decrease energy consumption. However, it has a negative impact on performance leading to Service Level Agreement (SLA) violations. Accordingly, selecting which VM to migrate from one host to another is a challenging task since it can affect performance. Researchers came up with several solutions and policies for efficient VM selection. In this paper, we exploit the fact that many tasks and users may tolerate some performance degradation which means, the tasks running on the VMs can be of different priorities. Accordingly, we propose augmenting consolidation with the priority concept, where low priority tasks are always selected first for migration. Towards this goal, we modified the popular Minimum Migration Time VM selection algorithm using the priority concept. The efficiency of the proposed algorithm is confirmed through extensive simulations using CloudSim toolkit and a real workload. The results show that priority awareness has a positive impact on decreasing energy consumption as well as maximizing SLA obligation.
Reliability is the biggest concern facing future extreme-scale, high performance computing (HPC) systems. Within the current generation of HPC systems, projections suggest that errors will occur with very high rates in future systems. Thus, it is fundamental that we detect errors that can cause the failure of important applications, such as scientific ones. In this paper, we have presented a two-level fault-tolerance approach for the detection and classification of errors for Compute United Device Architecture (CUDA)-based Graphics Processing Units (GPUs). In the first level, it detects the existence of errors by using software redundancy that applies design diversity. In the second level, it investigates the problematic software version and re-executes it on a different hardware component to classify whether the error is a permanent hardware error or a software error. We implemented our approach to run on GPUs and conducted proof of concept experiments by running three versions of matrix multiplications with different error scenarios and results show the feasibility of the proposed approach.
Even though the contribution of cloud computing towards the Sustainable Development (SD) of communities is still under research investigation, cloud computing has become an integral part of many ICT solutions that shape our daily lives. Thus, some researchers recommend taking considerable actions to point cloud computing development towards supporting SD. In this research, an approach to designing energy efficient cloud architecture as a way of supporting SD is proposed. Resource allocation is a challenging process in cloud management, the goal is to allocate the exact amount of resources needed throughout the service duration; tight enough to avoid unnecessarily wasting resources and loose enough to prevent any degradation in Quality of Service (QoS) that may lead to the violation of the Service Level Agreement (SLA) between the service provider and the cloud user. This study aims to achieve the desired balance by benefiting from the history of the user's behaviour and from sharing resourcesmore specifically Virtual Machines (VM)among a coalition of users. Coalition formation strategy is used to build groups of cloud users based on their cloud behaviour history. Users are grouped in a way that their usage patterns complement each other, either to avoid the loss stemming from VM excess reserved space or from idle times. A type of architecture that fulfils this improvement process is proposed and implemented on Google Compute Engine (GCE). The contribution of this research is that it applies the Coalition formation strategy in cloud computing resource management in a novel way and experiments show that there are scenarios where the efficiency of resource management has improved. Evaluation of the performance of the proposed architecture is done by comparing resource utilization for both the cloud following this architecture and the cloud that runs the basic GCE strategy. In conclusion, it is observed that improvements depend on accuracy of the prediction of usage pattern of the user. Results show that in certain scenarios, improvements can be made to up to 24% of VM usage and, in other scenarios, it can minimize the number of required VMs, thus contributing to green computing.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.