arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2001.07496cs.DC
收藏

开放联邦云环境下面向自组织Agent的动态资源分配

Self Organization Agent Oriented Dynamic Resource Allocation on Open Federated Clouds Environment

Kemchi Sofiane, Abdelhafid Zitouni, Mahieddine Djoudi

摘要:

为了确保联合云提供商向云客户提供不间断的服务,必须保证向用户高效分配云资源,以提高客户满意度和服务提供质量。最好是获得尽可能多的计算和存储资源。在云计算领域,已经提出了多种多Agent资源分配方法来实现动态资源分配问题。然而,这个问题仍然是开放的,在这一领域还有许多工作要做。在云计算中,鲁棒性是非常重要的,因此本文重点研究了开放联邦云计算环境下的自适应方法。我们的方法是混合的,我们首先采用现有的组织优化方法的自组织在代理代理组织,结合现有的多代理资源分配方法在联邦云。我们考虑一个开放的云联盟环境,它是动态的,并在不断发展,新的云运营商可以加入或离开这个联盟。同时,我们的方法是多标准,可以考虑到各种参数(即计算负载平衡的中介代理,地理距离(网络延迟)之间的客户和供应商.)。

英文摘要:

To ensure uninterrupted services to the cloud clients from federated cloud providers, it is important to guarantee an efficient allocation of the cloud resources to users to improve the rate of client satisfaction and the quality of the service provisions. It is better to get as more computing and storage resources as possible. In cloud domain several Multi Agent Resource Allocation methods have been proposed to implement the problem of dynamic resource allocation. However the problem is still open and many works to do in this field. In cloud computing robustness is important so in this paper we focus on auto-adaptive method to deal with changes of open federated cloud computing environment. Our approach is hybrid, we first adopt an existing organizations optimization approach for self organization in broker agent organization to combine it with already existing Multi Agent Resource Allocation approach on Federated Clouds. We consider an open clouds federation environment which is dynamic and in constant evolution, new cloud operators can join the federation or leave this one. At the same time our approach is multi criterion which can take in account various parameters (i.e. computing load balance of mediator agent, geographical distance (network delay) between costumer and provider...).

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑