Boolean feature spaces are prevalent in prominent domains such as spam detection, disease prediction, and sentiment analysis, yet their high dimensionality often limits classification accuracy and computational efficiency. To address these challenges, this paper introduces the Probability-Based Boolean Feature Selection (PBFS) algorithm, which adopts a step-wise, condition-based approach for fast feature selection, designed to evaluate Boolean features through variance and probabilistic class-dependent frequency measures. Across 29 test cases, PBFS achieves the highest number of top accuracy scores and the best mean rank among all evaluated dimensionality reduction methods. It significantly reduces classifier runtime while maintaining one of the lowest feature selection execution times. The 239 experimental outcomes highlight PBFS as a competitive, scalable, and computationally efficient feature selection method for high-dimensional Boolean data, offering a new perspective on feature relevance assessment that supports effective classification in complex domains.

Indoor localization is one of the most requested applications for 5G networks. With the continuous deployment of new 5G networks, the implementation of a reliable, robust, and accurate indoor positioning algorithm has become a widely and thoroughly researched topic. However, the solutions realized for real-world environments still do not meet all performance expectations. In this paper, a real-world 5G environment is investigated and the feasibility of the measurements collected for positioning is discussed. Based on these findings, a framework for indoor positioning is proposed and evaluated using Time Difference of Arrival (TDOA) measurements. Within the framework, an iterative algorithm is implemented to minimize a non-linear least squares function extended with four additional methods to mitigate the measurement errors caused mainly by non-line-of sight propagation. The best parameter settings are determined with hyperparameter optimization, and it is shown that with these settings, sub-meter positioning error is a viable goal in certain scenarios.

Estimating Quality of Experience (QoE) and Quality of Service (QoS) metrics is crucial for delay-sensitive use cases, and it relies on the available information in an operator's network. However, for low-latency traffic types that use unreliable and best-effort transport solutions for data transfer, obtaining these metrics only based on transport network probing – and ignoring the radio characteristics – is challenging. While end-to-end QoS metrics can be obtained by correlating per session radio and core data, predicting QoE is more difficult and requires the use of machine learning (ML) models. Nevertheless, training these models demands a vast amount of diverse measured reference metric data, which can be hard to acquire. In this study, we propose a delay-critical service-focused data collection framework that automatizes the measurements of networked services under various synthetic network degradation and jointly monitors radio and transport metrics, e.g., radio quality, throughput, latency, and user plane traffic. Our proposed framework runs on general-purpose laptops, eliminating the need for any specialized or expensive hardware. Furthermore, our tool provides a significant amount of data and meta-labels in an easy and automated way, which can be used to train future ML models to estimate QoS and QoE for the examined service. This demo paper demonstrates our proposed data collection framework, with synthetic transport and radio degradation, on a video conferencing use case and provides insight into how the main disturbance types influence the service streams.

Accurate indoor positioning is a highly requested feature in a wide range of applications, and an increasing demand is expected with the rollout of the fifth-generation (5G) cellular communication system. Although intensive research has been conducted on this topic since the 1990s, the global market still lacks a reliable, affordable, flexible, and sufficiently accurate indoor positioning system. Indoor localization over the cellular 5G network is a promising alternative for multiple reasons: no additional infrastructure costs, high availability, good control over security, and easy integration with other services due to standardization. The only questionable aspect is meeting the accuracy requirements for certain applications. However, 5G has numerous beneficial modifications and new features concerning positioning that can be exploited in the future. Our paper investigates the realization of the first TDOA-based indoor positioning system on existing 5G small cell networks, focusing primarily on the challenging effects of indoor signal propagation and possible ways to overcome them. To achieve this, we created a channel model based on real 5G measurements taken in an open-office building, and we used the obtained novel model to create a realistic simulation framework. This simulation framework was utilized to investigate the positioning performance of several algorithms. In addition to signal propagation issues, we investigate and highlight several other crucial aspects (e.g., synchronization and installation errors) that must be considered when deploying an industry-grade TDOA-based 5G positioning system.

With the advent of 6G and its mission-critical and tactile Internet applications running in a virtualized environment on the same physical infrastructure, even the shortest service disruptions have severe consequences for thousands of users. Therefore, the network hypervisors, which enable such virtualization, should tolerate failures or be able to adapt to sudden traffic fluctuations instantaneously, i.e., should be well-prepared for such unpredictable environmental changes. In this paper, we propose a latency-aware dual hypervisor placement and control path design method, which protects against single-link and hypervisor failures and is ready for unknown future changes. We prove that finding the minimum number of hypervisors is not only NP-hard, but also hard to approximate. We propose optimal and heuristic algorithms to solve the problem. We conduct thorough simulations to demonstrate the efficiency of our method on real-world optical topologies, and show that with an appropriately selected representative set of possible future requests, we are not only able to approach the maximum possible acceptance ratio but also able to mitigate the need of frequent hypervisor migrations for most realistic latency constraints.

Several recent studies shed light on the vulnerability of networks against regional failures, which are failures of multiple nodes and links in a physical region due to a natural disaster. The paper defines a novel design framework, called Geometric Network Augmentation (GNA), which determines a set of node pairs and the new cable routes to be deployed between each of them to make the network always remain connected when a regional failure of a given size occurs. With the proposed GNA design framework, we provide mathematical analysis and efficient heuristic algorithms that are built on the latest computational geometry tools and combinatorial optimization techniques. Through extensive simulation, we demonstrate that augmentation with just a small number of new cable routes will achieve the desired resilience against all the considered regional failures.

In this article, we propose a new protection scheme for backbone networks to guarantee high service availability. The presented scheme does not require any reconfiguration immediately after the failure (i.e., it is proactive). At the same time, it does not require any reserved backup network resources either. To achieve these seemingly contradictory goals, we utilize the recent advancements in Machine Learning (ML) to implement a network intelligence that periodically re-allocates the unused capacity as protection bandwidth to meet the service availability requirements of each connection. Our goal is achieved by two components (1) predicting the traffic for the next period on each link, and (2) intelligently selecting the best fit dedicated protection scheme for the next period depending on the estimated unused (spare) bandwidth and the previous service availability violations. Note that re-allocating protection bandwidth affects neither the operational connections nor the current best practice of operators to over-provision network bandwidth to support elephant flows. Finally, we provide a case study on the real traffic from Energy Sciences Network (ESnet), a high-speed, international scientific backbone network. The key benefit of our framework is that adaptively utilizing the over-provisioned bandwidth for spare capacity is sufficient to improve the availability from three-nines to five-nines (in ESnet for the 30 examined connections). The drawback is negligible bandwidth limitations; the user perceives a minor and very temporal bandwidth limitation in less than 0.1% of the time.

Generalized diversity coding is a promising proactive recovery scheme against single edge failures for unicast connections in transport networks. At the source node, the user data is split into two parts, and their bitwise XOR is computed as a third redundancy sub-flow. In order to guarantee instantaneous failure recovery without costly node upgrades, the network must ensure that any two of the three sub-flows reach the destination node in case of a single edge failure only by allowing flow duplication or merging identical flows, and avoiding any coding operation in the core network. In this paper, we investigate the corresponding routing problem to calculate capacity-efficient routes for these sub-flows. We propose a polynomial-time algorithm for topologies without capacity constraints on the links and without capability limitations of the nodes. We show that with node limitations the presented algorithm (as well as a minimum cost disjoint path-pair) provides a 4/3-approximation for the routing problem. Furthermore, we formulate an integer linear program to provide a minimum cost solution with arbitrary constraints in general graphs and we propose a polynomial-time algorithm in directed acyclic graphs. Our simulation results suggest that with upgrading only a small set of core network nodes with flow duplication and merging capabilities most of the benefits of generalized diversity coding can be achieved.

In this article, we investigate the Advanced Global Neighborhood Failure Localization (AG‐NFL) monitoring trail (m‐trail) approach, which provides ultra‐fast all‐optical restoration for any shared protection scheme. In contrast with its previous counterparts, AG‐NFL separates the management tasks of protection switching and link maintenance, and focuses on the identification of the proper switching actions in a timely manner rather than unambiguously localizing link failures. We form switching link groups at each node, that is, links whose failures do not have to be distinguished from each other, for example, because their corresponding switching actions can be performed at the same time. Forbidden link‐pairs are introduced to identify the minimal set of conflicting switching actions, which minimizes the number of switching link groups. Furthermore, in order to minimize the number of m‐trail reconfigurations upon dynamic traffic, we analyze the AG‐NFL performance in four different m‐trail design scenarios with decreasing dependency on the data plane. We prove that AG‐NFL is NP‐complete, and we propose an efficient heuristic to solve it. We demonstrate through simulations that unambiguous localization of switching link groups instead of single link failures leads to a significantly improved m‐trail performance both in wavelength resources and the number of required transponders, while signaling‐free restoration is still provided. © 2017 Wiley Periodicals, Inc. NETWORKS, Vol. 70(4), 327–341 2017

In this paper, we propose a new proactive recovery scheme against single edge failures for unicast connections in transport networks. The new scheme is a generalization of diversity coding where the source data AB are split into two parts A and B and three data flows A, B, and their exclusive OR (XOR) A⊕B are sent along the network between the source and the destination node of the connection. By ensuring that two data flows out of the three always operate even if a single edge fails, the source data can be instantaneously recovered at the destination node. In contrast with diversity coding, we do not require the three data flows to be routed along three disjoint paths; however, in our scheme, a data flow is allowed to split into two parallel segments and later merge back. Thus, our generalized diversity coding (GDC) scheme can be used in sparse but still two-connected network topologies. Our proof improves an earlier result of network coding, by using purely graph theoretical tool set instead of algebraic argument. In particular, we show that when the source data are divided into two parts, robust intra-session network coding against single edge failures is always possible without any in-network algebraic operation. We present linear-time robust code construction algorithms for this practical special case in minimal coding graphs. We further characterize this question, and show that by increasing the number of edge failures and source data parts, we lose these desired properties.

Survivable routing methods have been thoroughly investigated in the past decades in transport networks. However, the proposed approaches suffered either from slow recovery time, poor bandwidth utilization, high computational or operational complexity, and could not really provide an alternative to the widely deployed single edge failure resilient dedicated 1 + 1 protection approach. Diversity coding is a candidate to overcome these difficulties with a relatively simple technique: dividing the connection data into two parts, and adding some redundancy at the source node. However, a missing link to make diversity coding a real alternative to 1+1 in transport networks is finding its minimum cost survivable routing, even in sparse topologies, where previous approaches may fail. In this paper we propose a polynomial-time algorithm with O(|V||E| log |V|) complexity for this routing problem. On the other hand, we show that the same routing problem turns to be NP-hard as soon as we limit the forwarding capabilities of some nodes and the capacities of some links of the network.