The Internet of Things (IoT) enables the smart devices to be inter-connected. They share information with each other, with us and cloud based applications. These devices combine the physical and digital world and produce a huge amount of data to enhance the productivity of life, industries and society by providing smart services. IoT applications based on smart sensors open a new challenge which is the need of big data storage and huge computation power to provide real time data processing. IoT-Cloud solves such a problem since it provides a huge storage capacity. It also provides users on-demand access to resources at any place and any time.This work is designed to support any system where a huge data is generated and processed in real time such as a traffic monitoring system, a health system for obesity management using sensory and social data. We propose in this paper a new Infrastructure as a Service (IaaS) that provides an intelligent data storage to minimize the latency of any input and output data requests in a massive data storage and a huge number of servers. To ensure a high critical data availability, our IaaS supplies Cloud servers with high monitoring, backup and recovery services in case of a server failure. The proposed approach is termed Reliable lOad Balancing Using Specialization for ioT critical application (ROBUST). We compared the latency of an output file request and the complexity of searching the replicated version of a critical data of ROBUST to a recent IaaS architecture called Load Balancing in the Cloud Using Specialization (LBCS) and the classic one. The results shows a remarkable enhancement in terms of the complexity and the latency of an output file request.