Omid: Optimized Micro-Architecture Design For Fog Computing Applications

By the progress of Internet of Things (IoT) and the increasing number of connected nodes, new concepts such as edge and fog computing appeared to prohibit unneeded data transfer to reduce delay, bandwidth and power consumption. Although these concepts provide more processing, connectivity and storage capabilities to support different IoT use cases, but they also create new challenges especially in circuit and system domains. They require a hardware with configurable architecture, adequate data processing and network management capabilities which also considers high energy efficiency. Considering these requirements, this paper seeks to present a performance optimum processor architecture based on a specific systematic design methodology. Hence, an algorithm is proposed in this design methodology to design an optimum data-path based on the two proposed architectural solutions. This algorithm along with the micro-architectural optimization techniques, concludes the final optimum design among the feasible data-paths. The implemented benchmark applications on the final optimum architecture represents a significant improvement in performance with an acceptable power consumption and area overhead.