Caching through aerial platforms for 6g Networks

The staggering growth in IP Traffic, the development of ICTs (Information And Communication Technologies) and IoT (Internet of Things) devices and connections have led to an increase in the demand for new applications, and the need for a ubiquitous, reliable, and scalable service worldwide. These requirements cannot be entirely fulfilled by the 5G TNs (Terrestrial Networks) due to two main inherent limitations: the intrinsic energy costs of such an architecture, and its deployment in remote regions. Meanwhile, in addition to the aforementioned needs, the increased commercial viability of HAPS (High Altitude Platform Systems), made possible by the advances in solar panel efficiency, antennas, composite materials, and cost of manufacturing and launching, has fueled the conceptualization process of the next generation of wireless communication systems (6G). This has also led to the integration of aerial platforms in the overall network architecture, resulting in the creation of the so-called NTN (Non-Terrestrial Network) systems as an effective solution to complement TNs. According to the 3GPP (3rd Generation Partnership Project), NTNs, emerging as complementary to the current RANs (Radio Access Networks), are part of the so-called VHetNet (Vertical Heterogeneous Network). This new architecture, where the non-terrestrial part has a crucial role, is composed of three connected layers: a satellite network, an aerial network, formed by UAVs (Unmanned Aerial Vehicles) and HAPS, and a terrestrial network. This work focuses on the aerial segment, using a HAPS platform fulfilling the role of the above-mentioned layer, by equipping it with a MEC (Multi-access Edge Computing) server, thus providing it with caching capabilities. In this thesis, we assume that the HAPS is placed over the Milan municipality at an altitude of 20km and is designed to complement the TN in request handling, thanks to the offloading capabilities offered by the onboard MEC. Moreover, the HAPS offers coverage to 16 traffic zones within the municipality, whose traffic patterns are provided by a large Italian mobile network operator. These profiles report the value, in bits, of the traffic demand volume for 1420 BSs of Milan for two months in 2015 and they highlight districts' differences in terms of peaks and usage. In addition to that, for every zone, a sample cluster of BSs (Base Stations) is considered, comprising one macro-BS with ground caching capabilities, supported by six micro cell BSs that are not able to perform any content caching. This study has the aim of evaluating some key caching KPIs (Key Performance Indicators) such as the hit-and-miss probabilities and the quantity of traffic that has to be forwarded to the core network. Furthermore, it inspects how the integration of an aerial MEC changes the overall performance of some well-known caching algorithms and if some particular behaviors have to be noted when ground and non-terrestrial caches work together. For this reason, an in-depth analysis regarding the traffic profiles is provided, understanding their main features, and many possible scenarios are inspected. Results, carried out considering each of them, are obtained both in the case of single-beam HAPS coverage, meaning that the HAPS offers offloading capabilities to a single district of Milan, and in the multi-beam context, and they highlight that the considered parameters strongly vary from district to district depending on the traffic demand profile of the RAN portion under inspection.