Analytical resource planning for a multiservice node using an Erlang loss model, network slicing, and HAC/SAC control

This paper presents a multi-class access-node model - an Erlang loss system extended with resource-slicing parameters ((, ) and a global threshold for HAC/SAC admission control. Because the multiplicative property of the stationary distribution is preserved, we derive a linear recursive algorithm that computes blocking probabilities and resource occupancy within milliseconds, without exhaustive state enumeration. A comparative study of static versus dynamic slicing, combined with hard and soft admission control, shows that dynamic slicing with SAC increases channel utilization to ~95%, whereas a static slice with HAC minimizes blocking for priority traffic. The proposed method thus provides network operators with a fast, accurate tool for capacity planning in SDN/5G access nodes.