Analyzing Cost-Efficiency of Cloud Computing Applications with SimuLizar

Abstract

In cloud computing, software applications are potentially able to use only the computing resources that are minimally needed for performant operation. Because cloud providers provision such resources on a pay-per-use basis, software architects are interested in analyzing the operational costs that accrue for such applications, allowing architects to optimize for cost-efficiency. Current analysis approaches like Palladio focus on traditional performance metrics but lack support for cost-efficiency metrics. Therefore, software architects have to inaccurately estimate operational costs of their applications, potentially leading to economically unusable applications. To tackle this problem, we integrated cost metrics into SimuLizar, a Palladio extension for analyzing cloud computing applications. Our integration allows architects to attach prices to computing resources used by software applications. In a proof-of-concept evaluation with a simple book shop, we show that our integration allows architects to analyze costs with high accuracy.