Network-calculus-based analytical model for deterministic TDM-PON

Time-division multiplexing passive optical networks (TDM-PONs) are emerging as one of the most promising access network technologies for the industrial Internet due to their advantages of high bandwidth, low cost, and strong resistance to electromagnetic interference. Various applications within the industrial Internet, such as mobile robots and safety emergency stop applications, require TDM-PON to possess deterministic transmission capabilities. In response, a series of solutions, such as cooperative DBA (Co-DBA) and deterministic DBA (det-DBA), were proposed to enhance the deterministic network transmission capacity of TDM-PON. However, they were limited by focusing on some customized optimization problems but lacked a comprehensive analytical model to evaluate the inherent absolute performance upper bound under worst-case conditions and the practically guaranteed performance upper bound under specific scheduling strategies, which is crucial for describing determinism. To the best of our knowledge, we are the first to apply network calculus in TDM-PON to theoretically evaluate its performance. Our analytical model considers two fundamental bounds: the adversarial transmission window scheduling (A-TWS) bound, which describes the inherent performance bound under the most unfavorable scheduling, and the strategy-specific transmission window scheduling (SS-TWS) bound, which provides achievable performance guarantees under a particular allocation strategy. Based on this, we first established the arrival curve models for traffic flows and the service curve models for three bandwidth allocation schemes. Subsequently, we conducted a performance analysis of the delay and buffer size. Our results quantitatively compare the A-TWS and SS-TWS bounds of each bandwidth allocation scheme, intuitively contrasting the gap between the theoretical worst-case scenario and the practically achievable performance, which contributes to the performance evaluation of TDM-PON.