IoT Sensor Predictive Analysis System Using Apache Spark in Edge Cloud Environment

The Internet of Things (IoT) environment, which enables networking and computing in everything, is rapidly spreading. IoT environments cause bottlenecks and service delays to process data and provide services to users through a cloud-based central processing structure. In order to solve this problem, edge computing, which provides services to users by directly frosting data from IoT nodes and access networks, is rapidly spreading to cloud environment, where research continues to efficiently provide low delay intelligent services to users. In this paper, the response time of the service required to complete LSTM training is mainly measured by comparing to the locality of centralized data in the core cloud. In addition, the accuracy of the existing LSTM and the improved algorithm is compared on various reliability conditions.