A QoS-aware resources sharing architecture for homogeneous and heterogeneous wireless networks

The static model currently applied by governmental authorities for allocating the spectrum of frequencies and the increasing demand for network resources imposed by modern applications and services may lead to a resources scarcity problem in the near future. Dealing with this problem demands improvements on resources allocation. One of the ways of providing such improvements is by allowing resources sharing among network operators in both homogeneous and heterogeneous network scenarios. These network operators may implement different technologies, such as collective use of spectrum and licensed shared access to the spectrum of frequencies. Many related works have been proposed in the same context of the presented research, however these related works generally identify the need for additional resources and search for available resources without taking into account the QoS requirements of the resources renter and the costs involved in the resources sharing initiative. Therefore, in this thesis, a novel architecture is proposed to facilitate the implementation of resources sharing and consequently encourage network operators to lease their underutilized resources taking into account both the cost and the QoS requirements. This approach allows the network operator which is serving the resources to improve its profits at the same time that allows quality of service improvements to the resources renter. The main contributions of the proposed architecture include but are not limited to the design of a multilevel resources broker to control the resources sharing process. This broker is concerned on dynamically establishing a service level agreement that takes into account the quality of service requirements of resources renter. This process focuses on exchanging a small amount of control information to prevent the overhead from interfering with the legitimate traffic of the network operators. Another important contribution of the proposed approach is to improve the resources allocation in comparison with related work. Furthermore, the proposed solution is capable of taking fast decisions regarding resources allocation, what leads to the implementation of fast handover, allowing the traffic steering without interfering with incumbent users. The proposed architecture is modeled analytically and simulated using Matlab to evaluate its behavior in three different scenarios, considering both homogeneous and heterogeneous networks. The overhead in practical operation scenarios is kept under 1% of the total network traffic, what is considered not to interfere with the transmissions of the network operators. The fast decisions taken by the resources sharing architecture are based on accurate traffic load forecasting, what leads to fast handover, attaining times up to 46% lower than the maximum allowed handover duration. Results also show that both delay and jitter metrics are controlled to be maintained below their specific thresholds of the analyzed applications and therefore, the QoS is guaranteed for the resources renter, considering the coexistence of up to 500 devices.