Weighted Scheduling in Multipath GRE-in-UDP: Mathematical Modeling and Experimental Analysis

The growing complexity of modern network infrastructures and the proliferation of multi-interface devices have necessitated the development of advanced multipath communication solutions. Conventional TCP/IP networking, which relies on a single active path per session, is no longer sufficient to meet the demands of bandwidth aggregation, redundancy, and latency-sensitive services. Although Multipath TCP (MPTCP) partially addresses these limitations at the transport layer, it is constrained to TCP traffic and often requires substantial modifications to the operating system kernel. In this paper, we introduce a flexible, protocol-agnostic approach based on GRE-in-UDP encapsulation-termed Multipath GRE-in-UDP (MPT-GRE). We present a novel path selection algorithm that utilizes weighted scheduling to optimize traffic distribution across multiple paths. Our approach is thoroughly analyzed through mathematical modeling and validated with empirical data collected in controlled laboratory environments. The results confirm that MPT-GRE achieves high throughput, excellent load balancing, and minimal packet reordering, all while maintaining lightweight resource usage and broad compatibility across protocols and platforms.