Master of Science in Data Science (MSDS)
Network Slicing, Game Theory, Shapley Values, Incentivization
Next-generation communication networks are expected to support many different types of services. Network slicing virtually divides one physical network into several virtual networks in order to accommodate diverse demands for network connections. Assuming multiple network domains are stitched together to provide end-to-end network slices for such diverse services, it is critical to consider the economic behaviors and revenue models of the stakeholders. This project analyzes the economic behaviors from the perspective of a Global Slice Coordinator (GSC) that receives a slice request and coordinates the end-to-end resource allocations across multiple network domains. The primary role of the GSC is to incentivize each domain operator to report reliable statistics regarding each network domain, so that incoming requests can be mapped onto an appropriate set of resources that satisfy the service requirements, such as delay requirements, requested by slice users. We propose an incentivization mechanism based on the Shapley value, which divides the total revenue among the network domains based on the marginal contribution to the requirement satisfaction. We evaluate that the performance of our incentive mechanism succeeds in reducing the delay requirement violations and request rejections by obtaining more accurate results over time.
Mundhra, Yosha, "An Incentivization Mechanism towards Network-Slice as a Service" (2023). Master's Projects. 1340.
Available for download on Wednesday, January 01, 2025