An optimal storage utilization technique for IoT devices using sequential machine learning

Abstract

Internet of Things (IoT) is an emerging technology that integrates actuator devices with communication technologies for providing useful information exchange. The dimension of the actuator devices vary with respect to the application and environment it is deployed for. Small devices possess lesser operational capacity due to their limited memory space. Memory management in the devices is vital to improve the rate of processing and response of these devices. This article proposes a Selective Memory Balancing (SMB) technique for effective handling of memory to improve the rate of response of the actuator devices. The memory balancing scheme is built upon sequential machine learning algorithm that analyzes the periodic behavior of the device in handling requests. Based on the analysis, the available memory space is allocated and freed for receiving requests and storing information. The learning process is persuaded through time-dependent transmission behavior observations. This memory management technique operates in an adaptive manner for servicing time-critical and non-delay tolerant applications by minimizing storage and access delay at the device level. The proposed technique is intended to minimize memory exploitation, service delay and to increase the request processing rates.