An Approach to Steadfast and Reliable Resource Scheduling in Cloud Environment

Abstract

In cloud computing, resource allocation and scheduling are the necessary requirements which helps the cloud to access the resources that are available. Applications implemented in the cloud often required to interact amongst themselves and, in case, rarely depend on other implemented applications. Present cloud schedulers do not consider the infrastructure properties of cloud during resource allocation and scheduling process. In this paper, we propose a reliable resource scheduling algorithm that can automatically manage the cloud infrastructure by considering both user requirements and infrastructure properties and policies. A new prototype is presented based on Openstack and it implements the proposed cloud scheduler. Our work specifically focuses on providing the cloud scheduler with trustworthy input that expose the trust status of cloud infrastructure and it also further establish the foundations of planned future work to cover other properties. It further provides an implementation of the past work on trust management of cloud which provides the cloud I. INTRODUCTION Cloud infrastructure is generally complex and heterogeneous in nature which provides numerous components by various vendors. Applications implemented in the cloud often required to interact amongst themselves and, in case, rarely depend on other implemented applications. The rapid complexity of the infrastructure and application dependency leads an environment that requires management and raises security and privacy problems. The central component that restricts and controls the allocation of virtual resources for a cloud infrastructures physical resources is known as the cloud scheduler. Presently available schedulers do not consider security and privacy needs; neither do they consider the properties of the entire cloud infrastructure. So, a cloud scheduler should consider the necessary application performance requirements and also the users security and privacy requirements. This paper proposes a reliable resource scheduling algorithm that automatically manages the cloud infrastructure by considering both user requirements and infrastructure properties along with its individual policies. It also further develops the required trustworthy software agents that automatically manage the collection of the specific properties of physical resources. Maintaining a trustworthy and specific copy of the infrastructure properties and users requirements is critical for the absolute operation of the scheduler. This paper mainly focuses on providing the cloud scheduler with trustworthy input about the trust status of the cloud infrastructure which further establishes the foundations of planned future work. Open Stack refers to its specific cloud scheduler component using the term nova-scheduler and it identifies the cloud scheduler as the most complicated component to develop the significant effort still remains to have an appropriate cloud scheduler. In order to develop the trustworthy resource scheduler component, it is necessary to have the exact understanding of how clouds are managed and how they work in the environment. From this, it is finally concluded that establishment of trust in clouds requires two mutually dependent elements:  Appropriate trustworthy mechanisms and its tools to help cloud service providers activate the process of managing, maintaining, and securing the cloud infrastructure; and  Implementing methods for cloud and cloud providers for establishing trust in the operation of cloud infrastructure. This includes mainly supporting the cloud infrastructure with trustworthy and self-managed services that manages the entire cloud infrastructure. The automated self-managed services must provide the cloud computing environment with additional capabilities and more features. The example for additional capabilities are scale per use, hiding the cloud infrastructures complexity, automated true reliability, current availability, existing scalability, dependability, and resilience that considers the cloud security and privacy requirements by the properties.