Performance Evaluation of IoT-Fog-Cloud Deployment for Healthcare Services

Due to the proliferation of cloud computing and Internet of Things (IoT) systems in healthcare industry, remote real-time monitoring and tracking of patient conditions have become the norm. In a typical deployment, Medical IoT (MIoT) devices transmit patient data to a medical datacenters where the data get stored, analyzed, and made available to concerned parties or individuals. Healthcare providers, physicians and care takers can access remotely these medical data located in these medical datacenters which are typically deployed as a private cloud or public cloud. In such cloud deployment, the performance services to access the data in a timely manner is critical. One of the key performances metrics is the response time. It is important to deploy the medical datacenters in a cloud environment in a way to minimize the response time. Moreover, Fog/Edge computing layer is also used to perform localized processing and storage to further reduce response time. This paper studies, models, and shows how to reduce computing resources cost while guaranteeing health requests performance constraints-particularly response time of accessing medical data stored in a fog-cloud environment. To address this issue, this paper proposes a queuing model to predict the minimum required number of computing resources (both fog and cloud nodes) to meet the Service Level Agreement (SLA) for response time. We verify and cross validate our analytical model through discrete event simulation. Obtained results from our analytical models show that the proposed model is able to correctly and effectively predict the number of computing resources needed for health data services in order to achieve the required response time under different workload conditions.