arXiv subjects

Luiz DaSilva

Publications and source records attributed to Luiz DaSilva.

6 recordsLinked to original sources

arXiv2024

Decentralized Multi-Party Multi-Network AI for Global Deployment of 6G Wireless Systems

Multiple visions of 6G networks elicit Artificial Intelligence (AI) as a central, native element. When 6G systems are deployed at a large scale, end-to-end AI-based solutions will necessarily have to encompass both the radio and the fiber-optical domain. This paper introduces the Decentralized Multi-Party, Multi-Network AI (DMMAI) framework for integrating AI into 6G networks deployed at scale. DMMAI harmonizes AI-driven controls across diverse network platforms and thus facilitates networks that autonomously configure, monitor, and repair themselves. This is particularly crucial at the network edge, where advanced applications meet heightened functionality and security demands. The radio/optical integration is vital due to the current compartmentalization of AI research within these domains, which lacks a comprehensive understanding of their interaction. Our approach explores multi-network orchestration and AI control integration, filling a critical gap in standardized frameworks for AI-driven coordination in 6G networks. The DMMAI framework is a step towards a global standard for AI in 6G, aiming to establish reference use cases, data and model management methods, and benchmarking platforms for future AI/ML solutions.

cs.NI↗

arXiv2021

A Novel Open Set Energy-based Flow Classifier for Network Intrusion Detection

Several machine learning-based Network Intrusion Detection Systems (NIDS) have been proposed in recent years. Still, most of them were developed and evaluated under the assumption that the training context is similar to the test context. This assumption is false in real networks, given the emergence of new attacks and variants of known attacks. To deal with this reality, the open set recognition field, which is the most general task of recognizing classes not seen during training in any domain, began to gain importance in machine learning based NIDS research. Yet, existing solutions are often bound to high temporal complexities and performance bottlenecks. In this work, we propose an algorithm to be used in NIDS that performs open set recognition. Our proposal is an adaptation of the single-class Energy-based Flow Classifier (EFC), which proved to be an algorithm with strong generalization capability and low computational cost. The new version of EFC correctly classifies not only known attacks, but also unknown ones, and differs from other proposals from the literature by presenting a single layer with low temporal complexity. Our proposal was evaluated against well-established multi-class algorithms and as an open set classifier. It proved to be an accurate classifier in both evaluations, similar to the state of the art. As a conclusion of our work, we consider EFC a promising algorithm to be used in NIDS for its high performance and applicability in real networks.

cs.CR↗

arXiv2019

Power-Efficient Resource Allocation in C-RANs with SINR Constraints and Deadlines

In this paper, we address the problem of power-efficient resource management in Cloud Radio Access Networks (C-RANs). Specifically, we consider the case where Remote Radio Heads (RRHs) perform data transmission, and signal processing is executed in a virtually centralized Base-Band Units (BBUs) pool. Users request to transmit at different time instants; they demand minimum signal-to-noise-plus-interference ratio (SINR) guarantees, and their requests must be accommodated within a given deadline. These constraints pose significant challenges to the management of C-RANs and, as we will show, considerably impact the allocation of processing and radio resources in the network. Accordingly, we analyze the power consumption of the C-RAN system, and we formulate the power consumption minimization problem as a weighted joint scheduling of processing and power allocation problem for C-RANs with minimum SINR and finite horizon constraints. The problem is a Mixed Integer Non-Linear Program (MINLP), and we propose an optimal offline solution based on Dynamic Programming (DP). We show that the optimal solution is of exponential complexity, thus we propose a sub-optimal greedy online algorithm of polynomial complexity. We assess the performance of the two proposed solutions through extensive numerical results. Our solution aims to reach an appropriate trade-off between minimizing the power consumption and maximizing the percentage of satisfied users. We show that it results in power consumption that is only marginally higher than the optimum, at significantly lower complexity.

cs.NI↗

arXiv2016

Substitutability of Spectrum and Cloud-based Antennas in Virtualised Wireless Networks

Some of the new trends emerging in future wireless networks enable a vastly increased fluidity in accessing a wide range of resources, thus supporting flexible network composition and dynamic allocation of resources to virtual network operators (VNOs). In this work we study a new resource allocation opportunity that is enabled by the cloud radio access network architecture. In particular, we investigate the relationship between the cloud-based antennas and spectrum as two important resources in virtualized wireless networks. We analyze the interplay between spectrum and antennas in the context of an auction-based allocation mechanism through which VNOs can bid for a combination of the two types of resources. Our analysis shows that the complementarity and partial substitutability of the two resources significantly impact the results of the allocation of those resources and uncovers the possibility of divergent interests between the spectrum and the infrastructure providers.

cs.NI↗

arXiv2016

Energy and Spectral Efficiency Gains From Multi-User MIMO-based Small Cell Reassignments

In this work we investigate the reassignment of User Equipments (UEs) between adjacent small cells to concurrently enable spatial multiplexing gains through Multi-User MIMO (MU-MIMO) and reductions in energy consumption though switching emptied small cells to a sleep state. We consider a case where UEs can be reassigned between adjacent small cells provided that the targeted neighbouring cell contains a UE with which the reassigned UE can perform MU-MIMO without experiencing excessive multi-user interference, and whilst achieving a minimum expected gain in spectral efficiency over the previous original cell transmissions as a result. We formulate the selection decision of which UEs to reassign as a set covering problem with the objective of maximising the number of small cell base stations to switch to a sleep state. Our results show that, for both indoor and outdoor LTE small cell scenarios, the proposed MU-MIMO-based reassignments achieve significant reductions in the required number of active small cell base stations, whilst simultaneously achieving increases in spectral efficiency.

cs.NI↗

arXiv2015

Simulating Dense Small Cell Networks

Through massive deployment of additional small cell infrastructure, Dense Small cell Networks (DSNs) are expected to help meet the foreseen increase in traffic demand on cellular networks. Performance assessment of architectural and protocol solutions tailored to DSNs will require system and network level simulators that can appropriately model the complex interference environment found in those networks. This paper identifies the main features of DSN simulators, and guides the reader in the selection of an appropriate simulator for their desired investigations. We extend our discussion with a comparison of representative DSN simulators.

cs.NI↗