Queue Modeling for Geospatial Service on Edge-Cloud Architecture

The exponential growth and complexity of geospatial data necessitate innovative management strategies to address the increasing computational demands of GIS services. GIS is connected to the social context, and its use as a decision-support tool is gaining broader acceptance with the need to ensure high QoS. While cloud computing offers new capabilities for GIS, the physical distance between cloud infrastructure and end-users often leads to high network latency, compromising QoS. MEC emerges as a promising solution to limit latency and enhance system performance, particularly for real-time and multi-device applications. However, integrating GIS services into edge-cloud architectures presents significant challenges in terms of task scheduling and service placement. This paper proposes a queueing theory-based model designed to optimize the performance of GIS workloads within edge-cloud architectures. The model, based on a closed Jackson network, is designed to assist in the efficient design and deployment of edge systems that meet QoS and SLA requirements. The proposed framework is validated through a real-world case study, with performance metrics such as throughput and response time evaluated to ensure optimal system sizing and performance. The results underscore the potential of this approach for designing scalable and efficient edge-cloud architectures tailored to geospatial services.