To harvest idle, unused computational resources in networked environments, researchers have proposed different architectures for desktop grid infrastructure. The desktop grid system provide high computational capability at low cost and this motivates its use. However, there are several distinct differences between them in terms resource participation, resource sharing nature, application support, service quality, deployment effort required etc. Building a desktop system has to consider resource’s heterogeneity, non-dedication, volatility, failures, security, etc. Therefore, it is important to comprehend current research approaches in desktop grid system development to identify research gaps. In this paper, we propose a taxonomy for studying various desktop grid systems. We also present the strength and weakness of existing desktop grid systems and identified the research gap. © Springer International Publishing Switzerland 2016.