Tommaso Pecorella

Next-generation computing platforms are increasingly expected to accommodate a wide range of immersive applications, integrating advanced technologies such as extended reality and ultra-realistic virtual reality (uVR). The design and optimization of these distributed systems present significant challenges, particularly when managing the coexistence of traffic flows generated by both human and machine sources. This work introduces an end-to-end delay analysis framework focused on human-driven traffic in environments characterized by mixedsource flows. The human-generated streams-central to the analysis-coexist with machine-type communications that differ in terms of latency requirements. The delay analysis is carried out by deriving per-flow stochastic bounds, expressed in terms of the probability of receiving timely service. This is achieved through the use of stochastic network calculus and martingalebased methods, with human perceptual constraints explicitly incorporated into the theoretical model. Simulation-based validation confirms that the proposed bound accurately reflects actual end-to-end behavior. Moreover, the results demonstrate that integrating human perception into the analysis leads to improved performance compared to conventional approaches that do not account for cognitive aspects.

With the growing integration of renewable energy sources into distributed grids, accurate household - level load forecasting becomes essential for robust energy management and optimization. This paper proposes a lightweight stochastic profile generation method grounded in conditional probability approximation. First, empirical conditional distributions are mined from historical load data via hourly histogram binning and correlation analysis. Second, a Monte Carlo - inspired “flock” of plausible future load trajectories is generated iteratively, each endowed with an occurrence probability. Validation on the Ausgrid dataset (127 prosumer profiles over one year) shows that the probabilistic mining from an historical dataset of 30-180 days requires only 0.5-0.6 s, while generating 200 potential future scenarios takes 8.1 ms, with a total memory footprint of approximately 200 KB. These computational and storage efficiencies made the approach suitable for online deployment on edge devices, enabling robust optimization under uncertainty in renewable energy management.

As the Internet of Things (IoT) rapidly expands, there is a growing need for advanced routing solutions that accommodate constrained devices and diverse network requirements. Traditional single-metric routing approaches typically focused on parameters such as delay, hop count, or Maximum Transmission Unit (MTU), lack the adaptability required by IoT environments where energy efficiency, bandwidth considerations, and security constraints must be balanced. A more flexible protocol could improve network performance and prolong device lifetimes. To this aim, building on the extensibility of AODVv2, we developed and integrated a multi-metric routing protocol into the ns-3 simulator. Our implementation allows for the prioritization of multiple factors, such as residual battery levels, node reliability, and hop count, through a customizable metric ordering. Using Type-Length-Value (TLV) encoding, the protocol seamlessly incorporates additional metrics without compromising existing functionalities. Experimental results demonstrate that our multi-metric AODVv2 outperforms the original AODVv2 in terms of network performance, reduces energy consumption, and enhances the overall network longevity, thus providing significant benefits for IoT network management. Our released code paves the way for the programmability of other routing metric systems, including battery discharge rates, and the development of dynamic trust adjustment mechanisms for robust and adaptive routing.

High-Altitude Platform Stations (HAPS) are gaining recognition as effective communication platforms due to their rapid deployment capabilities and cost efficiency. Unlike terrestrial infrastructures, HAPS provide enhanced flexibility and scalability, making them well-suited for the development of Non-Terrestrial Quantum Wide Area Networks (NTQWANs). These networks overcome the inherent limitations of Optical Fibers (OFs), such as significant photon losses over long distances, which hinder efficient quantum communication. HAPS offer a distinct advantage over terrestrial networks by enabling wide-area coverage without the need for extensive ground infrastructure. Despite their benefits, the performance of Free Space Optic (FSO) links can be affected by atmospheric disturbances and positional variations in HAPS. To address these challenges, this paper explores an architecture for Wide Area Quantum HAPS Networks (WAQHNs) that supports applications such as Distributed Quantum Computing (DQC) and Quantum Key Distribution (QKD). Performance evaluation focuses on specific key metrics that are optical attenuation, fidelity, and entanglement rate, demonstrating that HAPS-based networks can outperform traditional OF-based systems.

The Quantum Internet (QI) necessitates a complete revision of the classical protocol stack and the technologies used, whereas its operating principles depend on the physical laws governing quantum mechanics. Recent experiments demonstrate that Optical Fibers (OFs) allow connections only in urban areas. Therefore, a novel Quantum Satellite Backbone (QSB) composed of a considerable number of Quantum Satellite Repeaters (QSRs) deployed in Low Earth Orbit (LEO) would allow for the overcoming of typical OFs’ attenuation problems. Nevertheless, the dynamic nature of the scenario represents a challenge for novel satellite networks, making their design and management complicated. Therefore, we have designed an ad hoc QSB considering the interaction between Digital Twin (DT) and Software-Defined Networking (SDN). In addition to defining the system architecture, we present a DT monitoring protocol that allows efficient status recovery for the creation of multiple End-to-End (E2E) entanglement states. Moreover, we have evaluated the system performance by assessing the path monitoring and configuration time, the time required to establish the E2E entanglement, and the fidelity between a couple of Ground Stations (GSs) interconnected through the QSB, also conducting a deep analysis of the created temporal paths.

Current literature has identified several issues with centralized mobility management in 5G-based Intelligent Transportation Systems (ITS), where 5G-enabled vehicles are expected to exhibit varying mobility patterns. So, distributed mobility management (DMM), where control plane mobility anchors work in a distributed fashion, is of keen interest to current researchers. Although existing DMM solutions, such as distributed mobile IPv6 (DMIPv6), distributed proxy MIPv6 (DPMIPv6), distributed mobility management software defined networking (SDN-DMM), distributed routing (D-Routing), and their derivatives appear to work fine in evolving 5G-based ITS, they cannot be applied directly. Further, to gauge their adaptability to different ITS scenarios, a thorough investigation of their performances is very much needed. Our key contributions in this paper include investigating DMM issues in 5G-based ITS scenarios, porting DMM schemes in 5G ITS, and developing a Markov model to evaluate handoff performance. The results show that DMIPv6 and DPMIPv6 outperform SDN-DMM and D-Routing in terms of end-to-end delay and signaling cost, with the latter increasing delay by 6-23% and signaling cost by 135% and 78%, respectively. However, SDN-DMM and D-Routing excel in packet loss ratio (15-20% lower), throughput (20-25% higher), and packet delivery cost (40% lower). In edge cloud flow mobility, SDN-DMM and D-Routing show better packet loss ratio as packet loss is caused by servers’ sensitive application delays rather than handoffs.

Machine learning (ML) has emerged as a compelling approach to identify attacks in network traffic security. Existing malware detection strategies often concentrate on specific facets, such as efficient data collection, particular types of malware, or handling data scarcity. While valid, these strategies typically overlook the potential for minimizing sample size, focusing instead on data augmentation. This work introduces a novel method to determine the minimum sample size necessary to achieve a specified accuracy level, measured by the F1 score derived from the confusion matrix. We focus on TCP header traffic data transformed into images through flow-splitting techniques for multi-class traffic classification. In addition, we introduce a diffusion model to generate new synthetic traffic images and show that our method outperforms existing techniques in terms of stability and predictability. This study also compares the effectiveness of synthetic image augmentation using Generative Adversarial Networks (GANs) and Denoising Diffusion Probabilistic Models (DDPM) in improving image recognition and classification accuracy.

The adoption of Extended Reality (XR) technology has been hindered by the need for high-bandwidth and low-latency networks to provide immersive experiences. Head-mounted devices used in XR are still heavy and not portable, limiting the potential of XR applications in various contexts. In this paper, we propose EdgeVerse, a multi-user XR system that leverages edge computing and extended Berkeley Packet Filter (eBPF) to offload computation, thereby reducing the dependency on high-bandwidth networks in support of XR applications. Our approach enables lightweight clients, such as devices with an edge browser, making XR more accessible to users. The key design of EdgeVerse focuses on offloading the connection and network synchronization of multiple XR users at the edge. We used an XDP bidirectional router that processes XR traffic faster to enhance user interaction, responsiveness, and immersive experience. To establish the practicality of our approach, we evaluate our results on a prototype that indicates improved response times and reduced latency with respect to baseline solutions.