اسحاق سلمانپورمهلباني

To support large and diverse Internet of Things applications, fog computing and edge computing technologies have been introduced. These technologies complement cloud computing and provide cloud-like services at the network edge, helping to reduce latency and increase real-time responsiveness. Edge-based servers can host cloud-based services, which leads to reduced network load and latency in user requested services. The enormous scale, geographical distribution, and heterogeneity of edge computing nodes make service placement in such infrastructure a challenging issue. Moreover, the different expectations of users and their mobility and dynamics, along with the characteristics of Internet of Things devices complicate the development problem further. Therefore, to achieve better performance, service migration must be conducted while considering the user's mobility in the environment and their changing requirements. In this study, service placement at the edge of the cloud has been studied with regard to the user's location. By studying similar work, we concluded that user movement prediction can enhance the service placement problem. To predict users' movement trajectory, we used the Kalman filter. With this approach, services are migrated to the appropriate server based on the user's future path before the user arrives at that server. This strategy results in reduced latency in responding to requests and improved service quality. eva‎luation results indicate that the proposed method reduces delay by an average of 3.96% for a 2 MB packet size and 6.26% for a 16 MB packet size. Additionally, it reduces costs by an average of 5.7% compared to similar works, and it also improves the packet loss rate by 2%.

محاسبات مه , محاسبات لبه , محاسبات ابري , مكانيابي سرويس , فيلتر كالمن , پيشبيني مسير حركت

Fog Computing , Edge Computing , Cloud Computing , Service Placement , Kalman Filter , Trajectory Prediction

Isaac Salmanpour Mahlabami

Marzieh Malekimjd