APRÈN. Portal de la producció docent del professorat de la UPC

The emergence of satellite constellations for telecommunications has notably increased in the recent years, enhancing the accessibility to space, and making satellite networks indispensable. The integration of Non-Terrestrial Networks (NTN) with Terrestrial Networks (TN) and the optimization of their management has become essential. This study focuses on addressing the uncertainties posed by this challenge, with particular emphasis on the unpredictability of traffic generation. In satellite operations, the focus on managing uncertainties has been mostly within Earth Observation (EO) applications, while in telecommunications scenarios, uncertainties (and specifically, those related to the unpredictable nature of traffic generation) have not been addressed as extensively. In this work, an analysis of the main uncertainty sources in satellite networks has been performed, and they have been aggregated into three main aggregated sources based on modeling similarities: satellite failure, link quality, and traffic model. These aggregated sources are described with a Weibull probabilistic distribution for satellite failure, bounded limits on the window duration for the link quality, and a generalized methodology for the integration of stochastic traffic models. In addition, an uncertainty management model has been designed. The proposed methodology for traffic generation is validated within an extensive framework that includes network-related metrics and constraints to optimize task scheduling. The steps for incorporating stochastic variables into the optimization process are described in detail, and a new optimization parameter is introduced: the schedule certainty. The impact of the schedule certainty is evaluated in an Internet of Things (IoT) scenario. Furthermore, the uncertainty model implementation and validation is performed in a constellation management tool for 5G/6G technologies, the Constellation Management System (CMS). The results compare the former CMS to the probability aware (P-aware) CMS. The findings show how the P-aware scheduling maintains a precise level of certainty consistent with operator-defined thresholds, providing greater certainty for the same performance metrics compared to the baseline CMS. This framework supports satellite operators in dynamically adjusting service coverage and system efficiency by means of stochastic information provided by the traffic model. To the best of our knowledge, this is the first time that such an analysis has been performed in the context of task scheduling for a satellite communication application. The insights gained from this study offer valuable contributions to optimizing operations in complex satellite networks.

Las comunicaciones dentro de una aeronave incluyen desde la interconexión de elementos de control y monitorización, hasta funciones de entretenimiento a bordo. Es por ello que cumplen una función fundamental dentro de la industria aeroespacial, necesitando cumplir severos criterios para la seguridad de la misma. En este documento se describen y simulan uno de los estándares actuales, AFDX (Avionics Full-Duplex Switched Ethernet) y su potencial sustituto, TSN (Time-Sensitive Networks). Dentro de TSN se han probado diferentes conformadores del tráfico, concretamente CBS (Credit-Based Shaper) y TAS (Time-Aware Shaper). El objetivo del proyecto ha sido documentar y modificar código de simulación, para una posible aplicación a docencia y analizar los resultados obtenidos, comparando AFDX y TSN. Para realizar esta comparación se ha hecho uso de simulaciones de tráfico. Se han comparado diferentes simuladores de redes disponibles, llegando a la conclusión que OMNET++ es el más adecuado. El escenario de simulación es el de un Airbus A380 en el contexto de vuelo de crucero. La comparación entre las diferentes tecnologías se ha realizado en términos de latencia, tamaño de las colas y utilización de los enlaces. Mediante este análisis se ha mostrado cómo las capacidades de una red TSN, son beneficiosas y mejoran las redes AFDX. Esto se debe a la división que se hace en TSN de los tipos de tráfico y cómo se asignan diferentes instantes de transmisión en función de la prioridad del tipo de tráfico.

Multiuser holographic communication services, supporting realistic and volumetric user's representations (i.e., holograms) integrated into shared 3D spaces, have the potential to become the main medium for communication, social interaction and collaboration in the near future. However, despite noteworthy advances in this field, key remaining challenges still need to be addressed to overcome limitations in terms of performance, scalability, and adaptability. This work departs from a worldwide pioneering end-to-end platform for holographic communications and extends it with novel contributions to allow for an increased performance in terms of delays and/or scalability, while reducing the overall resources usage. In this case, it proposes a new modular architecture for holographic communications that is capable of instantiating communication modules (i.e., Selective Forwarding Units or SFUs) on the Edge servers of interest. This not only allows the selection of the most powerful or closest Edge server to host a multiuser holographic communication session, but also to replicate multiple SFU instantiations to balance their load, based on different criteria (e.g., geographical distance or number of connected clients), in sessions with scalability needs. These extensions are fully compliant, and can leverage, emerging Cloud Continuum and 6G enablers, like European Telecommunications Standards Institute (ETSI) Multi-access Edge Computing (MEC) standard and Network-as-a-Service (NaaS) Application Programming Interfaces (APIs), respectively, thus paving the way for a more efficient and sustainable deployment over next-generation networks. Likewise, such devised extensions lay the foundation for supporting the Orchestration of other in-cloud communication and/or processing modules (e.g., Multipoint Control Unit or MCUs, or Remote Renderers) over the Cloud Continuum in next research phases. In addition, the platform has been extended and integrated with other technological enablers to ease its effective and even automated deployment in scenarios of interest, lowering the costs in terms of time, need for personnel, and infrastructure purchasing. Furthermore, new technological components and features have been developed, like the possibility to send pre-recorded holograms from players, a web-based and lightweight headless player, and a metrics monitoring subsystem (based on Prometheus and Grafana). Many of its components have been virtualized (Docker) and an automated deployment testbed, based on Ansible, has been put in place. Diverse (five) distributed scenarios, including diverse test conditions for each of them, have been defined and evaluated. The obtained results not only prove the benefits and advantages of the contributions of this work in terms of performance (delays, scalability, stability) and sustainability (less resources usage), but also confirm the validity of the developed testbed to conduct cutting-edge research in this relevant topic.

This master thesis focuses on bridging the gap towards facilitating novel spine-leaf¿based intra-data center connectivity controlled through the TeraFlowSDN controller and using the L2EVPN technology. The thesis provides a detailed overview of the project phases, from pre-documentation to implementation and experimental validation.

This bachelors degree thesis includes the design and development of a gateway between a 5G network and a Time-Sensitive Network (TSN). On the one hand, TSN is defined as a series of standards developed by the IEEE 802.1 group with the objective of achieving temporal determinism of the frames in Ethernet networks (wired networks), thus granting quality of service in industrial network scenarios, by means of the classification of flows depending on its criticality. This is performed thanks to the use of techniques like scheduling through the IEEE 802.1Qbv (Time Aware Shaper), the temporal synchronization of the involved devices (802.1AS), or the credit-based fair queueing mechanism (802.1Qav), among others. On the other hand, 5G is constituted as the fifth generation of mobile communication technologies. The improvements in bandwidth and latency compared to its predecessor, the 4G, turn it into a good option to develop new applications in the industrial and internet of things (IoT) environments, where parameters as the delay are critical, and provides mobility and flexibility, for instance, in the distribution of PLC (Programmable Logic Controllers) of the machines that are involved in the industrial process. The 3GPP organization is in charge of establishing the specifications regarding the 5G technology, and defines various ways of functioning of the aforementioned technology in the 5G-TSN integration, which can act as a cable or as a switch. Therefore, the objective of the project is to carry out an approximation that contributes to integrate a 5G network inside a network with TSN traffic, focusing on the time synchronization of the devices and the quality of service mapping from TSN to 5G. The proposed scenario in this bachelors degree thesis contemplates the use of hardware which is suitable to work in TSN networks, combined with emulators that will imitate the behaviour of a 5G network, both the user plane and the control plane.

Space has been populated by a wide range of satellite systems from governmental and private space entities. Monolithic satellites have been ruling it by providing a custom design that accomplishes a specific mission. Nevertheless, novel user demands emerged have required global coverage, low revisit time, and ubiquitous service. The possibility to integrate in-orbit infrastructure to support current communications systems has been discussed persistently during the last years. Specifically, the concept of deploying networks composed of aircraft and spacecraft (creating the so-called Non-Terrestrial Networks), has emerged as a potential architecture to satisfy this new demand. This novel concept has enabled to investigate mobile technologies in space infrastructure. For example, this is the case of the Software-Defined Satellite, which aims at managing in-orbit infrastructure by using Software-Defined Network techniques. These novel concepts pose multiple challenges which dedicated developments shall address. Likewise, specific equipment and simulation environments shall support them. Currently, open source satellite network emulators have certain limitations or are not easily accessible. This project aims at presenting the Virtual Satellite Network Simulator, a novel simulation engine capable to represent satellites as well as ground nodes in virtual machines and deploy a virtual network that depicts the channel effects and dynamics. VSNeS has been generated from different modules, that thanks to the joint work is able to generate the virtualization. First of all, a Python3 program has been developed, which works as a manager and is responsible for running the rest of the modules according to the virtualized scenario. Furthermore, Kernel-based Virtual Machine has been implemented for the execution of the virtual machines. The channel management is done with the NetEm emulator. Finally, a graphical user interface is delivered by Cesium. This dissertation presents formally a preliminary design with the essential steps to select each technology. Then, the networking design is also discussed. Different tests are also shown in order to verify the correct functioning of the tool. In addition, tests about the performance of the final release have been performed. The program has been tested with the following protocols in different realistic scenarios: TCP, UDP, and ICMP. This allowed us to verify the correct operation of the program, checking the delays and channel losses. Moreover, it is empirically demonstrated that some protocols are not functional for geostationary satellites, due to the long latency caused by the large distances.

This work aims at researching the existent solutions of monitoring and alerting techniques, as well as defining a suitable architecture, design and implementation of a complete and customizable monitoring and alerting framework used to inspect and notify specific conditions on dynamically instantiated applications operating in the network. Such Network Services (NS) are used in the Network Function Virtualization (NFV) architecture, allowing rapid instantiation and configuration of virtualized environments that handle network configuration. This design and implementation seek to provide more flexibility and dynamicity to the network operator to monitor custom or generic metrics and trigger notifications based on custom thresholds, without depending on the Virtual Network Function (VNF) developer to adapt its descriptor and onboard each version into the NFV Orchestrator (NFVO) prior to each usage. The framework here developed follows a modular architecture that separates the monitoring and alerting policies from the onboarding and instantiation process of the Network Functions. The architecture also facilitates the integration with other systems and adapting the functionality of an operational environment thanks to its decoupled and modular approach. The presented work considers a monitoring and alerting framework that is especially useful for dynamic environments such as those relying in NFV, like those in the EU H2020 PALANTIR project. There, the framework is used to help assessing the correct behavior of the Security NSs that are used to prevent or mitigate security anomalies in the network of each client. If abnormalities are found, remediation measures will take place to replace the potentially compromised NS instances with clean, appropriate ones.

Traditional satellite communications missions are based on artificial satellites that can communicate with ground stations. This type of network provides wide-area coverage as well as resilient communications. Satellite systems are expected to be merged with the introduction of 5G/6G mobile systems to offer seamless connection and ubiquitous coverage for users worldwide. SDN/NFV are two of the most important enabling technologies for deploying new 5G/6G mobile network architectures, and they are widely used in telecommunications to deliver different services. Its method is based on network softwarization, which abstracts the physical infrastructure by separating the functionality of the hardware. Furthermore, using SDN, it is possible to reconfigure the network in flexible topologies that adapt to the performance necessary at any given time, thereby maximizing resources and lowering costs. SDN/NFV, on the other hand, relies on a strong and continuous network layer to construct a control plane and deploy VNF. However, establishing a stable link is difficult in constellations of several LEO/MEO satellites. TALENT is a software platform created with the purpose of supplying satellite and ground connection by coordinating satellite and terrestrial systems from a single location, developed under the SaTG5 project that defined the integration of satellite solutions for 5G. Regardless of these solutions, a robust testbed is required to help in the discovery of novel protocols for delivering and orchestrating network services over satellite networks that take dynamic network architecture on satellite constellations into account. This thesis addresses the lack of a testbed for deploying VNF via a satellite network, as well as the orchestration and deployment of VNF across various satellite constellations. Its main contribution is an update to TALENT to allow it to work with newer Open Source MANO versions, and a testbed for deploying a VNF across a simulated satellite network based on the OpenSand satellite simulation tool.