Cloud computing is a flexible and scalable solution for delivering information technology (IT) services. However, as the processing power used must be paid for, accurately right-sizing the cloud resources for particular end user applications is important and can be expensive and time-consuming to determine. The authors present iCanCloud, a simulation platform for modeling and simulating public cloud environments that can provide accurate and repeatable performance evaluations of complex cloud-based scenarios.

Beginning with a short introduction of the issues, the design principles of the simulator are discussed and compared to existing cloud simulation tools. The features and design of the simulation platform are described in some detail and supported by appropriate diagrams and code examples. The hypervisor module that is the core of the simulator is covered, along with the central processing unit (CPU), memory, storage, and network systems. The usefulness of the simulator is then illustrated by implementing provisioning and scheduling phases for two cloud algorithms.

The correctness and accuracy of the simulator are then validated using Phobos, an application developed for the Finnish Mars MetNet space mission. The results of the validation processes are presented using a cost-per-performance metric and demonstrate a good level of agreement. Results of a favorable comparison of execution time and memory consumption with an existing mature simulation tool (CloudSim) are also presented.

The authors present a flexible and scalable tool for simulating public cloud environments that claims performance benefits and a simpler graphical user interface (GUI) compared to existing simulation tools.