IоT GATEWAY ARCHITECTURE FOR INTEGRATION WITH DIGITAL TWIN PLATFORMS

Abstract

This article investigates the quality of service (QoS) characteristics of an IoT edge gateway implemented on a resource-constrained Raspberry Pi platform in the context of digital twin systems. The study justifies the transition from a cloud-only approach to edge architectures to reduce latency, network load, and dependency on communication reliability. A prototype IoT gateway is designed and implemented to collect telemetry from sensor nodes (ESP32/Arduino), perform buffering and storage, and provide data access via an API.

A method for evaluating the impact of telemetry storage modes on QoS metrics is proposed, including storage latency, throughput, and packet loss. A controlled workload profile in the range of 1–10 Hz is generated using a stable telemetry source to ensure experimental consistency. The evaluation is based on statistical latency characteristics, including mean values, median, and upper percentiles (p95, p99).

A comparative analysis of two storage strategies—synchronous (sync) and batch—has been conducted. The results show that the sync mode ensures low and stable storage latency of approximately 1.6 ms across all tested frequencies. In contrast, the batch mode reduces the frequency of disk operations but increases latency to hundreds of milliseconds, with upper percentiles reaching approximately 1 second.

The findings demonstrate that the storage mode is a critical factor affecting the QoS of an IoT edge gateway and should be selected based on application requirements, such as minimizing latency or optimizing resource usage. The proposed methodology and experimental testbed can be used for further research and practical configuration of edge nodes in digital twin platforms.