NETWORK SLICING ALGORITHMS CASE STUDY: VIRTUAL NETWORK EMBEDDING

• Main Author: Irawan, Dedy
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/52255
• Institution: Institut Teknologi Bandung

In the 5G telecommunication network, one promising technique is network slicing. The network slicing technique enables infrastructure service providers to create end-to-end virtual networks from radio access networks to the core network. This virtual network consists of abstracted functions and resources. One of the network slicing issues is how to efficiently allocate virtual network resources on the substrate network. This can affect network performance in general. Resource allocation is strongly influenced by algorithms and computation time in mapping virtual networks into substrate networks and it is important to note because this affects service quality and profit for infrastructure service providers. From several studies conducted by the author, the problem of resource allocation in network slicing can be transformed into an optimization problem. The optimization problem in network slicing is known as virtual network embedding (VNE). In this report, the authors test the virtual network embedding algorithms of GRC, MCTS, and RL to compare profit gains for infrastructure service providers using long-term average revenue metrics and computation time in mapping virtual network allocations. It can be concluded that for profit the RL algorithm is 1% better than GRC and MCTS. Meanwhile, the computation time of the GRC algorithm is faster than MCTS and RL