Datasets

MIRAGE-COVID-CCMA-MA-2023

MIRAGE-COVID-CCMA-MA-2023 takes into consideration the traffic generated by more than 50 experimenters using 14 mobile apps for communication and collaboration via 2 devices. The experimenters used each app to perform at most 4 different user activities. In this new dataset, we analyzed the traffic when the user performs multiple activities (MA) during a single session with a given app (i.e., multiple activities in a single traffic capture).
The dataset will soon be available for downloading.

Download

MIRAGE-COVID-CCMA-2022

MIRAGE-COVID-CCMA-2022 takes into consideration the traffic generated by more than 150 experimenters using 9 mobile apps for communication and collaboration via 3 devices. The experimenters used each app to perform at most 3 different user activities. The dataset is released in two formats, making available both the raw traffic data captured (in JSON format) and a pre-processed version providing the set of inputs (in pickle format) leveraged in our work.

Download

MIRAGE-COVID-CCMA-2022 and MIRAGE-COVID-CCMA-MA-2023 datasets are released under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

CrowdSourced Collection

Photo by Windows on Unsplash

We collected our datasets by utilizing the Mirage architecture, which was conveniently optimized for capturing the traffic generated by communication and collaboration apps. The collection took place in the ARCLAB laboratory at the University of Napoli "Federico II".

Four different mobile devices were used for the experiments, all running on Android 10: a Google Nexus 6, a Xiaomi Mi 10 Lite, and two Samsung Galaxy A5.

The collection campaigns involved more than 200 volunteer students, both undergraduate and graduated, and researchers as experimenters who took on the role of typical users of the apps under consideration.

During each capture session, the experimenters engaged in a particular activity on a selected communication and collaboration app. This approach allowed us to obtain a traffic dataset that reflects the typical usage patterns of these apps. The duration of each session varied from 15 to 80 minutes, depending on the specific activity being performed.

Consequently, each session produced a PCAP traffic trace, accompanied by relevant ground-truth information extracted from additional system log-files. Using established network connections through the standard Linux command "netstat," the capture system reliably labeled each biflow with the corresponding Android package name. Furthermore, we enriched this information by adding a custom label that described the specific activity carried out by the user operating the device.

Mobile Apps

Discord

GotoMeeting

Meet

Messenger

Skype

Slack

Teams

Webex

Zoom

Clash Royale

Crunchyroll

JitsiMeet

KakaoTalk

Line

Signal

Twitch

User activities

Audiocall

Chat

Videocall

Standard

Methodology and Implementation

This section describes the methodology and implementation aspects of traffic classification and prediction for communication-and-collaboration apps via Deep Learning approaches. Further details can be obtained by clicking on "More Details".

Internet Traffic Classification - More Details

Internet Traffic Prediction - More Details

Dissemination

Publications

Fine-Grained Traffic Prediction of Communication-and-Collaboration Apps via Deep-Learning: a First Look at Explainability, Idio Guarino, Giuseppe Aceto, Domenico Ciuonzo, Antonio Montieri, Valerio Persico and Antonio Pescapè, IEEE International Conference on Communication (IEEE ICC), 2023.

In this work, we focus on four of these apps (Skype, Teams, Webex, and Zoom), whose traffic we collect, reliably label at fine (i.e. per-activity) granularity, and analyze from the viewpoint of traffic prediction. The outcome of this analysis is informative for a number of network management tasks, including monitoring, planning, resource provisioning, and (security) policy enforcement. To this aim, we employ state-of-the-art multitask deep learning approaches to assess to which degree the traffic generated by these apps and their different use cases (i.e. activities: audio-call, video-call, and chat) can be forecast at packet level. The experimental analysis investigates the performance of the considered deep learning architectures, in terms of both traffic-prediction accuracy and complexity, and the related trade-off. Equally important, our work is a first attempt at interpreting the results obtained by these predictors via eXplainable Artificial Intelligence (XAI).

Contextual Counters and Multimodal Deep Learning for Activity-Level Traffic Classification of Mobile Communication Apps during COVID-19 Pandemic, Idio Guarino, Giuseppe Aceto, Domenico Ciuonzo, Antonio Montieri, Valerio Persico and Antonio Pescapè, Elsevier Computer Networks, Special issue on Machine Learning empowered Computer Networks, 2022.

In this paper, we focus on communication and collaboration apps whose traffic experienced a sudden growth during the last two years. Specifically, we consider nine apps whose traffic we collect, reliably label, and publicly release as a new dataset (MIRAGE-COVID-CCMA-2022) to the scientific community. First, we investigate the capability of state-of-art single-modal and multimodal Deep Learning-based classifiers in telling the specific app, the activity performed by the user, or both. While we highlight that state-of-art solutions reports a more-than-satisfactory performance in addressing app classification (96%–98% F-measure), evident shortcomings stem out when tackling activity classification (56%–65% F-measure) when using approaches that leverage the transport-layer payload and/or per-packet information attainable from the initial part of the biflows. In line with these limitations, we design a novel set of inputs (namely Context Inputs) providing clues about the nature of a biflow by observing the biflows coexisting simultaneously. Based on these considerations, we propose Mimetic-All a novel early traffic classification multimodal solution that leverages Context Inputs as an additional modality, achieving F-measure in activity classification. Also, capitalizing the multimodal nature of Mimetic-All, we evaluate different combinations of the inputs. Interestingly, experimental results witness that Mimetic-ConSeq—a variant that uses the Context Inputs but does not rely on payload information (thus gaining greater robustness to more opaque encryption sub-layers possibly going to be adopted in the future)—experiences only F-measure drop in performance w.r.t. Mimetic-All and results in a shorter training time.

Classification of Communication and Collaboration Apps via Advanced Deep-Learning Approaches, Idio Guarino, Giuseppe Aceto, Domenico Ciuonzo, Antonio Montieri, Valerio Persico and Antonio Pescapè, 2021 IEEE International Workshop on Computer Aided Modeling and Design of Communication Links and Networks (CAMAD), 2021.

In this work, we focus on five such apps, whose traffic we collect, reliably label at fine granularity (per-activity), and analyze from the viewpoint of traffic classification. To this aim, we employ state-of-art deep learning approaches to assess to which degree the apps, their different use cases (activities), and the pairs app-activity can be told apart from each other. We investigate the early behavior of the biflows composing the traffic and the effect of tuning the dimension of the input, via a sensitivity analysis. The experimental analysis highlights the figures of the different architectures, in terms of both traffic-classification performance and complexity w.r.t. different classification tasks, and the related trade-off. The outcome of this analysis is informative for a number of network management tasks, including monitoring, planning, resource provisioning, and (security) policy enforcement.

Characterizing and Modeling Traffic of Communication and Collaboration Apps Bloomed With COVID-19 Outbreak, Idio Guarino, Giuseppe Aceto, Domenico Ciuonzo, Antonio Montieri, Valerio Persico, Antonio Pescapè, 2021 IEEE 6th International Forum on Research and Technology for Society and Industry (RTSI), 2021.

In this work, we address the characterization and modeling of the network traffic generated by communication and collaboration apps which have been the object of recent traffic surge due to the COVID-19 pandemic spread. In detail, focusing on five of the top popular mobile apps (collected via the MIRAGE architecture) used for working/studying during the pandemic time frame, we provide characterization at trace and flow level, and modeling by means of Multimodal Markov Chains for both apps and related activities. The results highlight interesting peculiarities related to both the running applications and the specific activities performed. The outcome of this analysis constitutes the stepping stone toward a number of tasks related to network management and traffic analysis, such as identification/classification and prediction, and modern IT management in general.

University Courses

Photo by Mikael Kristenson on Unsplash

Computer Networks

The topics of interest and objectives of the ADDITIONAL project are presented during the Computing Networks course, chaired by Prof. Antonio Pescapè at the University of Napoli Federico II for the Bachelor's Degree in Computer Engineering.

The Computing Networks course aims to provide the first theoretical notions and the necessary operational skills on computer networks and in general on packet switching communication networks. The course, following a top-down approach, favors an application vision of modern telematic technologies in the first place, to then arrive at the presentation of the software and hardware technologies at the basis of the creation of telematic systems.

During the Computing Networks course, the students are actively involved in the ADDITIONAL project, specifically in the collection campaign of the traffic generated by the communication-and-collaboration apps under study and its early characterization. The students are asked to carry out typical usage of these apps and perform specific activies to properly label collected traffic with both the app the activity tested.

Internet Data Analysis

The topics of interest and objectives of the ADDITIONAL project are presented during the course of Internet Data Analysis, chaired by Prof. Antonio Pescapè at the University of Napoli Federico II for the Master's Degree in Computer Engineering.

The Internet Data Analysis course aims to provide students with the specialized knowledge useful in analyzing a modern Internet network with special reference to network management and security aspects. The course presents the content by adopting an engineering and empirical approach and blends theoretical lectures, practical classes, seminars and exercises. It presents in depth the main aspects and motivations behind the analysis and performance evaluation of a network and then delves into the methodological and practical aspects related to network analysis with a specific focus on the analysis, identification and classification of anomalous events such as, for example, cyber attacks. The course also includes an exercise part functional to the development of a paper.

During the Internet Data Analysis course, the students are actively involved in the ADDITIONAL project, specifically regarding the analysis of the communication-and-collaboration apps' traffic via data-driven approaches, the production of technical reports, and the presentation of results to their colleagues.

DIGITA Academy

DIGITA Website

DIGITA is the "Digital Transformation & Industry Innovation Academy" of the University of Napoli Federico II in partnership with Deloitte Digital. It was created with the aim of providing young talents, with at least a Bachelor's degree, with the necessary skills to bridge the gap between companies and the Digital ecosystem, enabling them to be protagonists in digital transformation. The educational project consists of 9 months, 6 months of classroom lessons and 3 months of Project Work at partner companies.

The students and partner companies of the DIGITA were briefed on the subject matters and objectives of the ADDITIONAL project to foster its dissemination among both students with heterogeneous backgrounds and industry.

International Workshops and Conferences

Learn more about us

TRAFFIC is a research group of the DIETI Department at University of Napoli Federico II, working in the area of computer networks and multimedia, with focus on Network Monitoring and Measurements. Its research involves measuring and evaluating both experimental and operational systems/networks to derive knowledge and models for network behaviors.

TRAFFIC (and this is the main motivation of its name) focuses on monitoring Internet traffic to provide input for developing new algorithms, protocols, and systems for the current and future Internet, with a specific focus on network management and security.

Besides teaching activities, the members of Traffic are involved in several national and international research projects. They have a significant track record with respect to funding (approx. more than 5 Mln euro) and publications (e.g. conferences like SIGCOMM, Conext, IMC and journals like JSAC, IEEE Networks, IEEE Communication Magazines, etc).

We are part of the larger Computer Networks group (COMICS) at University of Napoli Federico II.