Application features‐based virtual machine deployment strategy in cloud environment

Abstract

Traditionally virtual machine (VM) deployment is based on one or two resources on the hosts and rarely considers other resources as well as the resource preferences of different applications. This leads to the imbalance and inefficient utilization of the resources on the physical hosts which further results in the waste of the cloud resources. To solve this problem, a novel VM deployment strategy is proposed. In consideration of the features of applications, the strategy deploys VMs based on the resource preferences of applications as well as the remained resources of the hosts. Utilizing the resource preferences the weights of various resources are determined, and the weight tables for VMs are obtained. Combining the resources weights of VMs and the resource configurations of physical hosts the values whether the resources are used in balance are calculated. To verify the correctness and effectiveness of the proposed strategy, extensive experiments were carried out on Cloudsim. Experimental results show that the proposed strategy is correct and effective. It outperforms other algorithms in VM deployment and ensures that the cloud resources are more efficiently utilized.

Performance Assessment of Using Docker for Selected MPI Applications in a Parallel Environment Based on Commodity Hardware

“…In the case of numerical integration tested in the paper, function f (x) = 1 1+x is to be integrated over the range [1,100]. The range is partitioned into a number of subranges which correspond to data packets and are initially distributed one per each slave process.…”

“…Nowadays, virtualized environments have gained much popularity thanks to ease of deployment and installation and are being used in various contexts such as: server/cloud deployment of applications provided by various parties possibly taking into account resource preferences of different applications [1]; IoT systems at various levels, i.e., server/cloud [2], middleware [3], gateway/fog [4] and sensor/edge [5] and, finally, in high-performance computing (HPC) systems potentially as a way to deploy applications in the cloud [6].…”