دانلود رایگان مقاله به اشتراک گذاری منبع بی سیم برای اپراتورهای متعدد

Abstract

Sharing wireless channel resources among cellular network operators saves cost by efficiently utilizing hardware, sites, and spectrum. To achieve high spectral efficiency, resource sharing has to adapt to the instantaneous load and channel state. However, before operators accept such dynamic scheduling policies, a solid understanding of their reliability and performance is required. In this paper, we present a consistent theoretical framework for Multi-Operator Scheduling (MOS). This formulation allows to analyze sharing guarantees and spectral efficiency for a large number of parameters and covers various fixed and dynamic resource sharing policies as special cases. Extending our study to the complete knowledge of the future channel state allows us to study the complete performance region of MOS. The proofs, analysis and simulation results in this work provide a profound understanding of the fundamental tradeoff between sharing guarantees and performance in multi-operator sharing. This insight will enable reliable yet efficient infrastructure sharing in 5G Radio Access Networks.

7. Conclusion

In this work, we proposed a consistent theoretical framework for wireless resource allocation to the users of multiple operators. Assuming a centralized scheduler at the BS, we formulated Multi-Operator Scheduling (MOS) as a convex optimization problem. By allowing the use of future channel states, we extended this formulation to Anticipatory Multi-Operator Scheduling (AMOS) and proved that MOS is a suboptimal case of AMOS. Both scheduling approaches allow to trade off sharing guarantees versus spectral efficiency by controlling the parameters maximum deviation and window size W. The rigorous analysis provided in this paper mathematically states the dependencies between these parameters and the system performance. We proved that the optimal utility monotonically increases with . In contrast, we proved that relaxing the constraint on W leads to a reduction of the utility with a probability that depends on the rate distribution. Simulation results confirm this analysis, characterize fairness among operators, provide insight into further interesting scenarios, and highlight the value of prediction for the AMOS formulation. We found that MOS and AMOS improve spectral efficiency without compromising fairness when operators have equal sharing ratios and users are equally distributed. The theorems and proofs in this paper provide a profound understanding of multi-operator scheduling. This new insight helps operators to understand the impact of resource sharing on their networks. Infrastructure providers can now offer flexible sharing agreements that match the sharing guarantees to the service requirements of a specific operator. This new degree in flexibility is particularly useful with the diverse set of operational parameters and service requirements in HetNets. In addition to this theoretical insight, our MOS formulation is highly practical, addressing the needs of the operators and having low computational complexity, and can be used to design sharing algorithms for today’s RANs without major architectural changes. It is our hope that this unique combination of theoretical insight and practical assumptions will lead to a fast implementation into future virtual Radio Access Network (RANs).