Meeting Hard Delay Constraint in Massive Access: A Mean-Field Approach

The emerging deterministic networking (DetNet) has stimulated an increasing enthusiasm for the investigation of supporting deterministic, ultra-reliable, and low-latency services. However, the time-varying channel characteristics and bursty data traffic bring uncertainties for transmission, thereby making the assurance of deterministic quality-of-service (QoS) a challenging issue in practice. In this paper, we are interested in supporting the deterministic QoS demand in a massive access scenario by reducing the bit dropping rate incurred by delay violation. To minimize the bit dropping rate, a cross-layer scheduling scheme with joint channel and buffer awareness is highly desired to efficiently adjust the resource allocation among users, whose complexity increases exponentially with the number of users. Fortunately, the complexity issue can be relieved by adopting the mean-field approximation approach, which can substantially simplify the design and analysis of the cross-layer scheduling scheme with massive users. Two threshold-based scheduling policies are proposed, which have low computational complexity. Besides, we also derive the deadline-constrained capacity for massive access, which is substantially superior to that of single user transmissions. Numerical results will demonstrate the effectiveness of the mean-field approximation based cross-layer scheduling scheme.