1,721,203 research outputs found
Foreword to the Special Issue on the 2017 Edition of the Workshop on Performance Evaluation of communications in DIstributed Systems and WEb-based Service Architectures (PEDISWESA 2017)
No full textPerformance evaluation is still a topic that attracts a lot of attention in both distributed and mobile systems as well as web-based services architectures. Due to the recent advances in Internet based applications as well as distributed and mobile communication systems, we are witnessing a variety of new technologies. However, these systems are becoming very large and complex at the same time. Several challenges remain to be resolved before these systems become a commodity. Guaranteeing Quality of Service (QoS) and effective provisioning of web-based systems as well as distributed and mobile systems as well as evaluating their communication performance still represent open issues in the design of these systems. Quantitative analysis can be very difficult and may be intractable because of the state space explosion
Understanding Human Mobility for CrowdSensing Strategies with the ParticipAct Data Set
Full text linkThe Mobile CrowdSensing (MCS) paradigm has been increasingly adopted in the last years. Its adoption has been proved as beneficial for different scenarios, such as environmental monitoring and mobility analysis. However, one of the major barriers of the MCS initiatives, is the difficulty in recruiting users for the purpose of collecting data. We focus in this work to such limitation, and we analyze the mobility traces collected with a real-world MCS experiment, namely ParticipAct. Our goal is to discuss how to exploit the mobility features of the recruited users, as grounding information to plan and optimize a MCS data collection campaign. In detail, we analyze the quality of the data set, its accuracy and several features of human mobility such as radius of gyration and the real entropy of the locations visited. We discuss the impact of such metrics on the task scheduling, allocation and how to obtain a certain Tcoverage of data from visited locations
Analysis of growth strategies in social media: The instagram use case
No full textSocial Networking describes the phenomena found in participatory and self-expressive Web sites-such as YouTube, Facebook and Instagram. Online communities represent a growing class of marketplace communities where participants can provide and exchange information on products, services, or common interests. Exploiting the phenomena, companies, artists, and new professional figures as influencers, youtubers and resellers, are increasingly using online communities to create value for their firms and customers, ensuring that their activities are relevant to the social network audience. However, from a marketing perspective, being able to distinguish a brand or an original work in a sea of competitors is a difficult challenge. From a more economic perspective, having a social media plan and strategy in place is becoming a must. At the same time, from the technical perspective, becoming popular in those platforms it is not so easy. This paper presents a seminal effort to investigate some strategies to growth on Instagram without using promotion tools. In particular, we studied how 'Benign' Social Bots, using the provided platform APIs, can affect the Social Media world. Our experimental results assess and benchmark the effectiveness of some approaches, increasing the diffusion of users content on the Social Media and consequentially speeding up growth
Accretion and ejection at work in the Narrow Line Seyfert 1 galaxy 1H 0323+342: A case of intermittent activity?
Full text linkContext. We present a comprehensive investigation into the properties of 1H 0323+342, a prominent jetted narrow-line Seyfert 1 galaxy. Aims. The primary objective is to understand the interplay between the relativistic jet, the hot corona, and the accretion disk around the supermassive black hole. Methods. This study spans the years 2006 to 2023, incorporating a rich dataset with 172 Swift observations, including the optical, UV, and X-ray bands, integrated with Fermi Large Area Telescope (LAT) observations. Spectral analysis was conducted on the X-ray observations using the XSPEC software, and the results were compared with optical, UV, and gamma-ray flux measurements and photon index values. Results. Our key findings include the identification of three distinct zones in the X-ray photon index-flux plot, characterized by high flux and a hard photon index (zone 1), high flux and a soft photon index (zone 2), and low flux and a soft photon index (zone 3). Before similar to 2017, 1H 0323 + 342 moved back and forth between zones 1 and 2; after that epoch, it transitioned to zones 2 and 3. Correspondingly, we observed a decreasing jet activity in the Fermi/LAT data and a reduction in the accretion rate in optical/UV data from Swift/UVOT. Conclusions. We interpret these observations in the framework of an intermittent jet scenario, driven by radiation-pressure instability in the accretion disk
Structured Sparse Ternary Compression for Convolutional Layers in Federated Learning
Full text linkIn Cross-device Federated Learning, communication efficiency is of paramount importance. Sparse Ternary Compression (STC) is one of the most effective techniques for considerably reducing the per-round communication cost of Federated Learning (FL) without significantly degrading the accuracy of the global model, by using ternary quantization in series to topk sparsification. In this paper, we propose an original variant of STC that is specifically designed and implemented for convolutional layers. Our variant is originally based on the experimental evidence that a pattern exists in the distribution of client updates, namely, the difference between the received global model and the locally trained model. In particular, we have experimentally found that the largest (in absolute value) updates for convolutional layers tend to form clusters in a kernel-wise fashion. Therefore, our primary novel idea is to a-priori restrict the elements of STC updates to lay on such a structured pattern, thus allowing us to further reduce the STC communication cost. We have designed, implemented, and evaluated our novel technique, called Structured Sparse Ternary Compression (SSTC). Reported experimental results show that SSTC shrinks compressed updates by a factor of x3 with respect to traditional STC and with a reduction up to x104 with respect to uncompressed FedAvg, at the expense of negligible degradation of the global model accuracy
The rhythm of the crowd: Properties of evolutionary community detection algorithms for mobile edge selection
Full text linkThe Multi-access Edge Computing (MEC) paradigm increases the computational capabilities of distributed sensing architectures, such as Mobile CrowdSensing platforms, which are designed to collect heterogeneous data from the crowd by exploiting mobile devices. In this context, our work focusses on the impact of three community detection algorithms to our edge selection strategy. In particular, we study TILES, Infomap, and iLCD which are specifically designed to identify evolving communities of users in dynamic networks. Our analysis is based on the ParticipAct data set that offers real human mobility data. We first measure the quality of the data set during an observation period of 1 year, during which the data set provides the 75% of the expected traces collected by approximately 170 users. We then compare some structural properties of the communities detected, namely Similarity, Forward Stability, Cohesion and Coverage. We conclude our study with a performance analysis of the selected Mobile MECs by varying the community detection algorithms adopted. In particular, we measure the latency and the number of satisfied requests and we show that the average latency obtained with Infomap is slightly lower than that of the other algorithms, while the average number of satisfied requests is higher when we adopt the TILES algorithm
Toward Fog-Based Mobile Crowdsensing Systems: State of the Art and Opportunities
Full text linkMCS is an emerging paradigm that leverages the pervasiveness of mobile, wearable, and vehicle-mounted devices to collect data from urban environments for ubiquitous service provisioning. In order to manage MCS application data streams efficiently, a scalable computing infrastructure hosting heterogeneous and distributed resources is critical. FC, as a geo-distributed computing paradigm, is a key enabler for this requirement as it bridges cloud servers and smart mobile devices. Research on the integration of MCS with FC has recently started to be explored, recognizing the requirements of MCS and their coexistence with cyber-physical systems. In this article, we analyze the state of the art of FC solutions in MCS systems. After a brief overview of MCS, we emphasize the link between MCS and FC. We then investigate the existing fog-based MCS architectures in detail by focusing on their building blocks, as well as the challenges that remain unaddressed. Our detailed review on the subject results in a taxonomy of FC solutions in MCS systems. In particular, we highlight the node structures, the information exchanged, the resource and service management, and the type of solutions adopted concerning privacy and security. Moreover, we provide a thorough discussion on the open issues and challenges by reporting useful insights for researchers in MCS and FC
A social-driven edge computing architecture for mobile crowd sensing management
Full text linkThe multi-access edge computing (MEC) architectural model has fostered the development of new human-driven edge computing (HEC) frameworks that extend the coverage of traditional MEC solutions leveraging people roaming around with their devices. HEC is a well-suited architecture for human-centered technologies such as mobile crowdsensing (MCS) as it allows conveying and distributing sensing tasks at the edges of the network, also enabling (local) sensing data collection from devices. This article, through the joint use of HEC and MCS paradigms, introduces a new social-driven edge computing architecture based on incentives and centrality measures. The core idea is to add social MEC (SMEC) nodes to complement the traditional edge nodes (i.e., the main actors of the middle layer of the standard MEC architecture), acting as bridges between other devices and the cloud. The principle that underlies the SMEC selection is based on the attitude of the users in performing tasks and on their measures of centrality. In addition, we report extensive experimental results based on co-location traces and cooperativeness scores extracted from the ParticipAct living lab, a well-known MCS dataset based on data collected between 2013 and 2015 from 170 students of the University of Bologna, that show how the selection based on centrality measurements returns greater benefits than simple selection based on cooperativeness scores
A Probabilistic Model for the Deployment of Human-Enabled Edge Computing in Massive Sensing Scenarios
Full text linkHuman-enabled edge computing (HEC) is a recent smart city technology designed to combine the advantages of massive mobile crowdsensing (MCS) techniques with the potential of multiaccess edge computing (MEC). In this context, the architectural hierarchy of the network shifts the management of sensing information close to terminal nodes through the use of intermediate entities (edges) bridging the direct Cloud-Device communication channel. Recent proposals suggest the implementation of those edges, not only employing fixed MEC nodes, but also opportunistically using as edge nodes mobile devices selected among the terminal ones. However, inappropriate selection techniques may lead to an overestimation or an underestimation of the number of nodes to be used in such a layer. In this article, we propose a probabilistic model for the estimation of the number of mobile nodes to be selected as substitutes of fixed ones. The effectiveness of our model is verified with tests performed on real-world mobility traces
- …
