Performance evaluation of single-tier and two-tier cloudlet assisted applications

As the usage of the mobile devices increases, the user expectations for the performance of the smart applications grow as well. One of the basic criteria of the performance is the seamless user experience. Edge Computing approaches are becoming popular for supporting real time applications for which the traditional cloud based solutions do not suffice. For time-sensitive applications such as the real time speech-to-text conversion, face recognition and simultaneous translation, users require fast responsivity while communicating with the remote servers. High latency during this communication will yield to poor user experience. To overcome this major bottleneck, the cloudlet which is a specific form of Edge Computing bring opportunities. The cloudlets provide better performance by faster service compared to the conventional remote cloud connections. The cloudlets are used for different purposes on different architectures to solve domain specific problems such as caching, reducing delay, taking real-time decisions and bringing the cloud resources to closer. In this paper, the performance of three possible cloudlet-assisted Edge Computing architectures are evaluated by considering a generic delay-intolerant application.