Using Multicast Network Slices in Edge Environments

Communication networks have undergone significant transformations by offering new services, traffic growth and increased demand for service quality, such as delay reduction, reliability, security and availability. Edge computing seeks to bring data storage and service processing closer to users, reducing delays and increasing the availability and reliability of the offered services. New technologies such as network virtualization through the use of Software-Defined Network (SDN) and Network Function Virtualization (NFV) offer greater ease of management and agility for adapting networks and offering new services and resources, while network slicing, which allows resource reservation, isolation among services and better reliability, seeking to meet the demands of new applications and guarantee the required Quality of Service (QoS). However, it is still challenging to meet peak demand in places with many users, such as sporting and cultural events and some public buildings. This work proposes using Multicast Network Slices (MNSs) to meet the demands of multimedia transmissions and other services that require resource reservation in edge networks in locations with a large concentration of users. In this paper, we describe the formalization of the multicast distribution tree discovery problem using Integer Linear Programming (ILP) and an algorithm for deploying MNSs with resource reservation in an edge network. The proposed algorithm was tested using as input one of the edge networks of the Brazilian Chamber of Deputies, and a set of broadcast, multicast and unicast network slice requests, and the results showed an average bandwidth savings of over 80% on core links and over 73% on distribution links in every test.