A congestion control framework for heterogeneous storage structure IoT node

Congestion control schemes are essential to optimize the transmission performance of internet of thing nodes. Under heterogeneous storage structures including single and hybrid storage structures, how to design and evaluate congestion control schemes remains an open question. Aiming at this problem, buffer scheduling probability distribution vector (BSV) is designed for quantitative describe result of inter-buffer scheduling under hybrid storage structures. Congestion control probability distribution vector (CCV) is designed to characterize single-buffer congestion control result. Based on BSV and CCV, a single-node performance evaluation framework with flexibly configurable parameters under heterogeneous storage structures is proposed. The framework capable of obtaining closed-form steady-state solutions for the comprehensive performance evaluation metrics of single nodes such as rejection rate, throughput and delay is proposed. Based on prediction and triage of self-similar network data, a expectation congestion control (PT-ECC) scheme and a probability distribution congestion control (PT-PDCC) scheme aims to improving node’s transmission performance is presented. Numerical results show that the PT-ECC scheme and PT-PDCC scheme can improve node’s performance and suppression effect for data-intensive suddenly traffic congestion compared to existing congestion control schemes. Furthermore, the effectiveness of the proposed system in evaluating and design congestion control schemes based on hybrid storage structures is demonstrated.