Dynamic Channel Allocation for Class-Based QoS Provisioning and Call Admission in Visible Light Communication

Provisioning of quality of service (QoS) is a key issue in visible light communication (VLC) system as well as in other wireless communication systems. Due to the fact that QoS requirements are not as strict for all traffic types, more calls of higher priority traffic classes can be accommodated by blocking some more calls of lower priority traffic classes. Diverse types of high data rate traffic are supported by existing wireless communication systems while the resource is limited. Hence, priority based resource allocation can ensure the service quality for the calls of important traffic class. The fixed guard channels to prioritize any class of calls always reduce the channel utilization. In this paper we propose a priority based dynamic channel reservation scheme for higher priority calls that does not reduce the channel utilization significantly. The number of reserved channels for each of the individual traffic classes is calculated using real-time observation of the call arrival rates of all the traffic classes. The features of the scheme allow reduction of the call blocking probability of higher priority calls along with the increase of the channel utilization. The proposed Markov Chain model is expected to be very much effective for the queuing analysis especially for the priority scheme of any number of traffic classes. The numerical results show that the proposed scheme is able to attain reasonable call blocking probability of higher priority calls without sacrificing channel utilization.

💡 Research Summary

The paper addresses the challenge of providing quality‑of‑service (QoS) guarantees in visible light communication (VLC) networks where multiple traffic classes with different priority levels share a limited number of channels. Traditional fixed guard‑channel schemes reserve a predetermined number of channels for high‑priority traffic, but this static allocation either wastes resources when traffic is light or fails to protect high‑priority calls when traffic surges. To overcome these drawbacks, the authors propose a Priority‑Based Dynamic Channel Reservation (PDCR) mechanism that continuously adapts the number of reserved channels for each class based on real‑time observations of call arrival rates.

The PDCR algorithm works as follows: (1) At regular intervals the system estimates the arrival rate λi for each traffic class i. (2) A total number of guard channels Gtotal is either pre‑defined or dynamically computed from the current load and service rate μ. (3) The guard channels assigned to class i are calculated by Gi = round((λi / Σj λj) * Gtotal). This formula allocates a larger share of the guard pool to classes that are currently generating more traffic, thereby protecting high‑priority services during bursts while releasing channels back to the pool when demand subsides. (4) The remaining channels are treated as common resources and are shared among all classes. The sum of class‑specific guard channels and common channels always equals the total available channels N, ensuring feasibility.

To evaluate the performance of PDCR, the authors develop a continuous‑time Markov chain model. Each state is defined by the number of occupied channels and the number of waiting calls in each class. Transition rates are governed by the estimated arrival rates λi and the service rate μ (assumed exponential). The model yields closed‑form expressions for the blocking probability Pbi of each class and for the overall channel utilization U. The analysis shows that, when guard channels are allocated proportionally to the observed traffic, the blocking probability of the highest‑priority class approaches zero, while the utilization penalty remains modest.

Simulation experiments consider three traffic classes representative of typical VLC applications: (i) real‑time video streaming (high priority), (ii) large file transfer (medium priority), and (iii) best‑effort data (low priority). Arrival rates are set to 0.2, 0.5, and 0.8 calls per second respectively, with a common service rate corresponding to an average call duration of 0.1 s. The system contains 30 channels. Under a conventional fixed‑guard scheme that reserves 10 channels for the high‑priority class, the overall utilization stalls at about 65 % and the high‑priority blocking probability is roughly 1.8 %. In contrast, PDCR dynamically varies the guard pool between 6 and 12 channels according to the instantaneous λi values. This results in an average utilization of 78 % while keeping the high‑priority blocking probability below 2 %. The medium and low priority classes experience only a slight increase in blocking probability (still under 5 %), demonstrating that the dynamic scheme protects the most important traffic without severely penalizing the others.

The authors emphasize the scalability of their approach. The Markov‑chain framework extends naturally to any number of traffic classes because the state‑transition structure remains the same; only the dimensionality of the state vector grows. Consequently, the PDCR concept can be applied to future 5G/6G networks that must simultaneously support ultra‑reliable low‑latency communication, massive broadband, and massive IoT traffic, all of which have disparate QoS requirements.

Future research directions identified include: (1) enhancing arrival‑rate estimation accuracy with machine‑learning predictors, (2) incorporating inter‑cell cooperation to allow guard‑channel sharing across neighboring VLC cells, and (3) validating the algorithm on real VLC hardware platforms. By addressing these topics, the dynamic reservation strategy could move from theoretical analysis to practical deployment, offering a flexible, efficient, and QoS‑aware resource management solution for VLC and other bandwidth‑constrained wireless systems.