Nutshell: Cloud Simulation and Current Trends
Ubaid ur Rahman,Owais Hakeem,Muhammad Raheem,Kashif Bilal,Samee U. Khan,Laurence T. Yang
DOI: https://doi.org/10.1109/smartcity.2015.51
2015-01-01
Abstract:Cloud computing has experienced enormous popularity and adoption in many areas, such as research, medical, web, and e-commerce. Providers, like Amazon, Google, Microsoft, and Yahoo have deployed their cloud services for use. Cloud computing pay-as-you-go model, on demand scaling, and low maintenance cost has attracted many users. The widespread adoption of cloud paradigm upshots various challenges. The legacy data center and cloud architectures are unable to handle the escalating user demands. Therefore, new data center network architectures, policies, protocols and topologies are required. However, new solutions must be tested thoroughly, before deployment within a real production environment. As the experimentation and testing is infeasible in the production environment and real cloud setup, therefore, there is an indispensable need for simulation tools that provide ways to model and test applications, and estimate cost, performance, and energy consumption of services and application within cloud environment. Simulation tools providing cloud simulation environments currently are limited in terms of features and realistic cloud setups, focus on a particular problem domain, and require tool-specific modeling, which can be frustrating and time consuming. This paper aims to provide a detailed comparison of various cloud simulators, discuss various offered features, and highlight their strengths and limitations. Moreover, we also demonstrate our work on a new cloud simulator "Nutshell", which offers realistic cloud environments and protocols. The Nutshell is designed to diminish flaws and limitations of available cloud simulators, by offering: (a) multiple datacenter network architectures, like three-tier, fat-tree, and dcell, (b) fine grained network details, (c) realistic cloud traffic patterns, (d) congestion control strategies and analysis, (e) energy consumption, (f) cost estimation, and (g) data center monitoring and analysis. Flexibility to stretch the architectures to simulate smart city IT infrastructure.