Pso Research Papers - Academia.edu (original) (raw)

—Cloud computing is a ubiquitous platform that can be used to access services and resources. The requirements for accessing a cloud resource is minimal, however, the users of a cloud platform are initially required to configure the type of services they need to access in the cloud. Even though cloud platform is elastic, frequent upscaling is costly. This paper presents an effective technique that can be used to automatically identify user's requirements and to allocate appropriate packages depending on the requirements. Usage logs of the client is grouped geographically and group based requirements are identified. This helps to determine the resource requirement for each group which address the problem of underutilization or overutilization of resources. These requirements are passed to PSO, along with the packages offered by multiple cloud operators in the same region to identify the best package for the current requirement. Experiments conducted on this architecture proves the effective working nature of the system in minimal time and with low QoS difference. Keyword-MCDM, cloud package selection, Optimization, PSO, multi-cloud I. INTRODUCTION Adopters for cloud platform have grown enormously due to the increase in the requirements for automation of systems. However, satisfying them has prevailed to be a controversial issue. This is due to the inappropriate selection of resources, leading to either underutilization or overutilization of resources. Cloud computing is a distributed computing technique that presents the user with an infinite pool of resources that can be accessed with minimum requirements [1]. Cloud resources are not available as such for the users to access them. Instead, packages are provided to the user with varied configurations such that the users can select a configuration of their choice and the resources are assigned accordingly [2]. Packages are usually segregated as high performance and high storage, moderate performance and high storage, low performance and low storage etc. Several such combinations are provided to the user with varying levels of quality parameters. The quality parameters include performance, storage, reliability, robustness, availability etc. These packages are defined by the cloud service providers and are to be used as such, without any modifications [3]. The process of package selection plays a vital role in determining the satisfactory level of the user. The major difficulty faced during this stage is the identification of the appropriate package that suits the current requirements of the user [4]. This problem can actually be branched into two sub-problems namely; identifying the current requirements of the user and selecting the appropriate package that suits the user's requirements. Both these issues are to be performed manually. This acts as the major drawback of the selection process. Automation of these two processes has not been considered by the cloud providers, hence cloud users who are technically not well versed face this issue to a large extent. Inability of a user to identify their service requirements leads to selection of inappropriate packages and this in-turn leads to dissatisfaction. In-order to overcome the issue of dynamic service requirement peaks, cloud servers incorporate the concept of elasticity by scaling up resources as and when required [5] [6]. The scaling process is automatically done, hence the user need not worry about the sudden increase in requirements. However, these automatic process scaling tends to be costly [7]. If these resource scaling spells are rare, they need not be bothered. However if they are frequent, a package change is usually recommended as obtaining a package with higher requirement is economical compared to frequent resource scaling [8]. As of now, identifying the appropriate package is a trial and error process, which starts with an educated guess that follows with the usage levels. However determining it automatically will lead to improved reliability of the cloud services [9]. This paper proposes a requirement identification and a package selection technique that can be used to dynamically identify the user's requirements and recommend appropriate packages corresponding to the requirements.