2021 RoEduNet Conference: Networking in Education and Research

4 Nov 2021, 12:00 → 6 Nov 2021, 14:00 Europe/Bucharest

Iasi

20th RoEduNet Conference
Networking in Education and Research

Nowadays, modern education and research activities are strongly dependent on a high-speed communication infrastructure and computer networks based on the newest technology. Design, implementation, management of such networks, together with development of new application fields are not possible without good knowledge of networking state-of-the-art.

The 20th edition of Agency ARNIEC/RoEduNet's (Romanian Education Network) annual Conference is organised with the help of "Alexandru Ioan Cuza" University of Iasi and IEEE Computer Society Romania Chapter, under the patronage of Ministry of National Education and Scientific Research of Romania, offers special opportunities for information exchange in computer networking: technical and strategic aspects, communication issues, and of course their applications in education and research.

Your experience and ideas are very interesting for us and for all the participants. You are welcome to participate. To stay informed about the 20th RoEduNet Conference, please create a user on conference website.

The conference will be organized online with help from our partners and sponsors.

Thank you, and hope to see you soon at the 20th Conference of RoEduNet !!!

2021 Call for Papers.pdf IEEE Templates

Thursday 4 November

13:30 → 15:30 Opening Plenary Session

Networking in Education and Research

https://uaic.webex.com/uaic/j.php?MTID=m65f81ea93992bb5e3caf7a7eda759314

13:30 Opening

Nicolae Țăpuș - IEEE România Chapter
Florin Bogdan Manolache - Carnegie Mellon University

14:00 Dell in mediul universitar: o poveste de succes - Valentina Frangu (Public Sector Lead, Dell Tech.)

14:20 Securing the Higher Education and Campus with Fortinet Security Solutions - Adrian Danciu (Fortinet Senior Regional Director South Eastern Europe)

14:40 Collaborative research & innovation to support tech transfer - Marius Iordache (Orange)

15:00 Why 400G ? - Catalina Niculita, Cisco Romania

15:30 → 16:00 Coffee Break

16:00 → 18:00 Partners - Technical Session

16:00 Accelerate innovation with HPC and Cloud by Dell - Daniel Tanase (Solutions Architect Dell Tech.)

16:40 Solutii de securizare a accesului in retea pentru mediul Educational Universitar - Dan Gabor (Senior Systems Engineer Fortinet)

17:20 Flexible Engine - Cristian Turcin (Orange)

Friday 5 November

09:00 → 10:40 Networking in Education and Research

Chair Paul Gasner
Networking in Education and Research Virtual Room A

Virtual Room A1

09:00 New developments of the RENAM-GEANT communication platform for support science and education in Moldova

The article describes the development trends of the national and regional networking infrastructures in Moldova and in Eastern Europe for effective support of research and educational activities. Special attention is focused on description of implementation of new Cross Border Fibers (CBF) connections, other components of the communication infrastructure in the framework of the UE EaPConnect project. Created networking infrastructure is the basis for deployment and operation of national e-Infrastructures and wide range of services offered by RENAM-GEANT platform.

Speaker: Dr Petru BOGATENCOV (RENAM Association)

New developments of the RENAM-GEANT_RoEduNet Conf_ver1.pdf

09:20 An SD-WAN Approach for EUt+ Network

Abstract— The European University of Technology (EUt+) is an alliance of eight universities which aims to offer students the possibility of studying at different partners all over the continent and obtaining a recognized diploma. Our goal is to propose a framework for the European University of Technology network, which needs to support both the physical and virtual mobility of students between the partners. We rely on Software-Defined Wide Area Network (SD-WAN) and Management and Orchestration (MANO) for deploying the network. Next, several traffic engineering mechanisms are added to find the best path for different types of traffic and thus achieve the highest Quality of Experience for the users. Finally, we employ homomorphic encryption to secure not only student and staff personal information but also to maintain the autonomy of each partner when it comes to handling private data.

Speaker: Dr Iustin-Alexandru IVANCIU (Technical University of Cluj-Napoca)

Ivanciu RoEduNet 2021.pdf

09:40 Online school - challenging the coronavirus pandemic

Abstract - The coronavirus pandemic has caused a crisis that affects several sectors of society, prepared to deal with critical situations. This is also the case of education, faced with a new challenge: the digital one. The aim of this survey is to evaluate the satisfaction and acceptance rate of online education among teachers and students at University of Medicine and Pharmacy “Carol Davila” from Bucharest, to find ways to improve the education system, both in times of crisis and in general. We conducted the survey by using the questionnaire as a tool, being a method that allows rapid collection of data from a large population. The online questionnaire was distributed to teachers and students form this university using social platforms, such as WhatsApp and Facebook. The survey was accessible for 2 weeks. After the data were collected through the online survey, they were analyzed, and the description of the answers was performed. The research sample consisted of a number of 52 teachers and a number of 239 students. The responders who answered the questionnaire are part of the following age categories: 13.46% (n = 7) in the category 21-29 years, 36.53% (n = 19) in the category 30-39 years, 32.69% (n = 17 ) in the category of 40-49 years and 15.38% (n = 8) in the category over 50 years. Over half of the respondents are university assistant professors (n = 29), a quarter (n = 13) lecturers, 11.53% (n = 6) associate professors and 5.76% (n = 3) university professors. This study showed that, despite issues such as poor internet connectivity or access to it, online courses have been attended by students and have proven to be a very pragmatic and feasible teaching option and can certainly complement traditional teaching. in class. So, we can deduce that soon enough people will get used to the online teaching environment, and blended learning that includes both online and offline teaching will be part of the standard curriculum. Another conclusion of the study is that the success of distance learning involves a collaborative effort between teacher, computer scientist, communication network performance and student.

Speaker: Mr Andrei NECSULESCU (Department of Emergency, „Dr. Carol Davila” Central Military Emergency University Hospital)

#_17_Online_school_final_4_IVN_presentation.ppt

10:00 Modelling academic responsibility in relation to student satisfaction for online courses at the "Dunărea de Jos" University of Galaţi, Romania – a proposal

Modelling academic responsibility in relation to student satisfaction for online courses at the "Dunărea de Jos" University of Galaţi, Romania – a proposal

Assist. PhD. Candidate Leonard GARABET
Prof. PhD. Luminița DUMITRIU
"Dunărea de Jos" University of Galaţi, Romania

Abstract
The paper proposes a conceptual model for the evaluation of the triad academic responsibility - quality of service – student satisfaction at University "Dunărea de Jos" in Galati, Romania in the context of online learning on the Microsoft Teams platform.
Starting from the concept of corporate social responsibility (Corporate Social Responsibility - CSR) the concept of academic social responsibility was derived, including the responsibility of universities in terms of teaching, but also the one related to the social and economic context of student training. CSR is generally understood as a strategic initiative that contributes to a brand's reputation.
Two quality criteria of the services were identified, evaluated, and then used, namely the quality of the E-learning services and the quality of the IT services offered by the university. These criteria have been studied in conjunction with student satisfaction, and the data collected and working hypotheses were processed with the SPSS application.
Over the existing model at this level we want to extend the model with an academic responsibility.
The model attempts to express the relation that should exist between the quality of service, the academic responsibility and student satisfaction and support for the university brand, either by his return to university, for lifelong learning, or by publicity.

Keywords: E-learning, User analysis, IT services, academic responsibility

Speakers: LEONARD GARABET (UNIVERSITATEA "DUNĂREA DE JOS" DIN GALAȚI), Prof. LUMINITA DUMITRIU (UNIVERSITATEA "DUNAREA DE JOS" DIN GALAȚI)

10:20 viztop – Intuitive Visualization of Remote Real-Time Monitoring of Linux Processes

We introduce here our prototype system for intuitive visualization of remote monitoring of the dynamic of processes in a running Linux operating system, in real-time. Such a system provides for overcoming the limitations of the text-based process monitoring commands or tools available in current Linux distributions. Thus, it can display, from anywhere in the Internet, via a web browser, the processes that exist at any given time in the running operating system using a graph of interconnected nodes (both processes and threads). Visual cues are used for representing different information about processes and threads, such as shapes, colors, size, text, lines, etc. These cues allow a large amount of information to be shown to the user, in a much easier way to understand when compared to classic text based tools from Linux (ps, top, etc.). As processes are created, terminated, or they change state, the nodes in the process graph are added, removed, or change shape, color, size, line type, and so on. There is a strong correlation between the visual elements and the characteristics of the monitored processes. The relations between processes are also shown. This system can be used as a tool that provides both the big picture with regard to resources’ usage in a computer system and offers plenty of details to be used for improving system administration. It can be also a valuable educational tool for students, helping them to understand the dynamics of processes in operating systems.

Speaker: Zoran CONSTANTINESCU (Petroleum Gas University of Ploiesti)

prez-viztop-roedu-2021-final.pptx

09:00 → 10:40 Social Aspects of Networking Environment Today

Chair Emil Cebuc
SANE Virtual Room B

09:00 IoT Enabled Optimized Architectures for GPS Anti-Theft Tracking Devices

In this paper we summarize in a split direction state of the art survey – marketwise and research-oriented and we present a twofold IoT enabled, anti-theft system architecture – one oriented towards bicycles, that do not have a battery or its own electrical system and one oriented towards electrical vehicles such as electrical bikes/scooters and cars. We corroborate the hardware architecture with a functional software architecture, that can be easily used on both solutions. In the end, we present the relevant results of a case study implementation of the architectures: the tracking map resulted from continuous monitoring of the GPS position and periodical transmissions by GPRS connection and the variations in SMS receiving times.

Speakers: Mr Alexandru Viorel PĂLĂCEAN (University POLITEHNICA of Bucharest, Computer Science and Engineering Department), Dumitru-Cristian TRANCA (University Politehnica of Bucharest)

09:20 Stamina - Assistive Technology platform for pandemic prediction and crisis management

Abstract—Crisis management can be defined as how an organization, or worldwide society, deals with an unexpected event that threatens to harm or even destroy it. The management of a crisis can be split into three phases: pre-crisis, where prevention activities take place; crisis response, in which the management sector has to deal with the problem; and post-crisis, where people try to come back to what they knew as an everyday life before the appearance of the crisis. Over the years, the healthcare domain has been dealing with various epidemiological cases (measles outbreaks in 2018, Influenza during 2018-2019, H1N1 pandemic crisis, etc.) having different levels of complexity. Internet of Things (IoT) solutions such as wearables, real-time analysis using smart systems have been developed to improve the patients' experience when in need and doctors' performance in the field. Still, very few complete systems for decision support, pandemic prediction, and management that connect people from different countries are available. STAMINA will deliver an intelligent decision support toolset for pandemic-related real-life situations, covering most healthcare systems' gaps. In the development process, STAMINA will use a combination of pre-existing technology not currently used by health emergency planners in their daily practice of pandemics management. The method involves gathering data to predict potential threats, assess the impact on financial and societal levels, and recommend mitigation actions.
Index Terms—healthcare, crisis management, pandemic, Internet of Things (IoT), early warning platform

INTRODUCTION
The demand for decision support systems for pandemic prediction and management in the human healthcare domain has increased due to the effects [1] of the Coronavirus disease 2019 (COVID-19). Significant deficiencies regarding countries' epidemic preparedness have been observed since the pandemic has started.
In this scope, the continuous evolution of the Internet of Things (IoT) solutions can improve the healthcare analysis and treatments regarding various diseases [2] by speeding the process and decreasing the amount of work for healthcare staff members (doctors, nurses, etc.). In return, the worker's stress level will decrease, the efficiency will increase because of the reduced number of daily tasks, and the mental health disorders among patients and healthcare staff will be prevented [3].
The Internet of Medical Things (IoMT) includes sensors and smart applications [4] that can examine the patients' condition and give feedback in real-time [5], supporting decision making and even collecting the healthcare parameters for future observations. The issue of an efficient early warning system for epidemiological diseases symptoms started to be of great interest in researchers' activities, primarily because of the benefits it can bring to the healthcare system and society. Biosensing systems are used in early notification and diagnosis approaches for faster detection of various diseases [6].
Collaborative crisis management tools play an essential role in consolidating country-to-country communication by creating solid relationships from which faster response strategies will arise. Furthermore, critical threats and challenges for each sector could be overcome through collaborative risk analysis and early planning [7].
This paper will present the STAMINA solution, which addresses all the improvements mentioned above, needed for better results in dealing with health emergencies (Influenza, SARS-CoV-2, Measles, West Nile Virus, etc.). The STAMINA toolset will include: real-time analysis of web and social media pages, wearable devices (portable diagnostic included), predictive modeling of a pandemic outbreak and its impact (along with decision-making support in implementing mitigation strategies), machine learning-based early warning system (ML-based EWS), crisis management tool, and preparedness pandemic training tool. The STAMINA vision has been designed considering the user perspective, with five main objectives: provide stakeholders with novel, easy-to-use software tools that complement EU-level systems; create a set of guidelines and best practices to improve preparedness and response; increase diagnostic capability; enhance cooperation between and within the EU Member States and neighboring countries, and finally, ensure the sustainability of the STAMINA solution [8].
The paper is structured as follows: Section II will analyze existing state-of-the-art, Section III will focus on describing the components, methods, and models of the STAMINA solution, Section IV will present the Romanian pilot requirements and status. Finally, Section V draws the conclusions and envisions future work.

STATE OF THE ART
After the pandemic influenza (H1N1) in 2009, the 2011 Escherichia Coli outbreak in Germany, the Ebola virus in 2014, Zika virus in 2016 and West Nile virus in Southern and Eastern European countries in 2019, the Coronavirus pandemic had a massive impact on humanity, forcing scientist and researchers from all over the world to implement innovative solutions in order to help prevent the virus from spreading at such a high rate.
(As mentioned before), STAMINA is currently developing an intelligent decision support tool-set for pandemic prediction and management and is demonstrating its efficiency by practitioners at national and regional levels within and across EU borders. On a similar note, Faculty of Economic Sciences and Business Management from Babes Bolyai University in Cluj has recently developed a research project that proposed a specialized online platform [9] through which the researchers of the university involved in this scientific initiative publish a series of relevant data on the economic impact of the COVID-19 pandemic in the form of interactive infographics, meant to show a comprehensive, updated real-time image of the Romanian economy. The main objective of the project is to provide real support to decision-makers in Romanian politics and economy by carrying out and regularly updating the analysis of the situation generated by the COVID-19 pandemic.
As it has become a common fact that people are not really relying anymore on old-school news sources to inform themselves about the epidemic situation, but rather on the Social Network Service (SNS), an study conducted by the Department of Computer Science and Engineering from Seoul National University of Science and Technology proposes a SNS Big Data Analysis Framework for COVID-19 Outbreak Prediction in Smart Sustainable Healthy City, based on information retrieved from the Twitter platform. The obtained results have demonstrated an outbreak cluster predicted seven days earlier than the confirmed cases. The possibility of analyzing data from SNS platforms enabled the prediction of outbreaks some days earlier, and eventually, the infection rate was reduced [10].
One study [11] conducted in China presents a mortality risk prediction model for COVID-19 (MRPMC) that uses patients’ clinical data in order to separate patients into different groups by mortality risk, which allows the prediction of physiological deterioration and death 20 days earlier. This proposed model is based on four machine learning methods including Logistic Regression, Support Vector Machine, Gradient Boosted Decision Tree, and Neural Network. It enables precise mortality risk stratification of patients with COVID-19, and facilitates more responsive health systems that are responsive to critical cases of COVID-19 patients.
Maybe one of the most popular safety measures that was taken in order to reduce the consistent number of COVID-19 cases was lockdown. All over the World, people were advised to stay inside of their home for as long as possible and keep the legal social distance between other people while being out in public. Even though these measures have been shown to considerably reduce the new confirmed cases, an article [12] proposes an objective and quantitative way to monitor population behavior in order to analyze the impact and response of people in COVID-restricted situations. The team of scientists explored the utility of the RADAR-based platform to test the effect and response of lockdowns and social distancing measures aimed at reducing the spread of COVID-19 by interpreting participant data already collected from November 2017 onward as part of the ongoing RADAR-CNS studies. The study concluded with some very relevant results: the team managed to quantify expected changes in time spent at home, distance travelled, and the number of nearby Bluetooth-enabled devices between the period before lockdown and the one during the lockdown. They also noticed reduced face-to-face gatherings and communication as measured through mobility features and increased virtual socialization through the phone, along with some health parameters of patients such as heart rate, which resulted in being not too high, and the chaotic sleep schedules of patients, caused by the tremendous pandemic.
The Coronavirus pandemic was a critical situation for humanity, forcing people everywhere, in a very unpleasant manner, to create new and innovative solutions in order to help not only themselves, but also people in need from all over the globe to successfully defeat the virus.
SNCRR - through the Department for Emergency Situations - consisting of volunteers and employees - is mandated to provide support to the authorities in the process of managing humanitarian crises - of any kind and to intervene to remove their effects. At the same time, SNCRR is mandated to participate in European research projects, encouraging innovation in the organization - and at the same time offering feedback and validation of the concepts of the participating projects. In the last year, SNCRR was also involved in the COVINFORM Project - one of the 23 new research projects funded by the European Commission with a total of 128 million euros to address the coronavirus pandemic and its effects [13].
Policymakers and public health experts unanimously recognise the disproportionate impacts of COVID-19 on vulnerable persons: even in countries with well-developed responses, the outbreak and its repercussions imperil the basic well-being of social groups whose livelihoods are already precarious.
COVINFORM draws upon intersectionality theory and complex systems analysis in an interdisciplinary critique of COVID-19 responses on the levels of government, public health, community, and information and communications. Promising practices are evaluated in target communities through case studies spanning diverse disciplines and vulnerable populations.
The COVINFORM project will: assess COVID-19 responses in a multilevel governance framework and develop an online portal and toolkit for stakeholders in the governmental, public health, and civil society/community domains.

COMPONENTS, METHODS AND MODELS
Infectious diseases have the potential to pose serious threats to public health. Managing this type of crisis remains a serious challenge due to the number of people involved, the different legal, administrative, professional and political cultures, as well as the lack of cross-border crisis management infrastructures.
STAMINA contributes to overcoming these challenges by providing improved decision-making technology for pandemic crisis practitioners at regional, national and European level. The project will focus on two stages of the emergency management cycle: preparedness and response.
This will be possible with the following tools:
Antimicrobial Resistance Model (AIR): currently focuses on the E.coli microbe along with one of its resistance mechanisms called ESBL; will expand its project to present x conditions of multitasking learning model (MTL) architectures to predict output estimates (e.g., number of deaths caused by prolonged symptoms and mortality rates) over a period of time.

Lifelong physical activity modeling and simulation (PALMS): it is used to model the effect of different policy interventions on long-term public health in response to an outbreak of COVID-19 in Valencia. PALMS can be used both independently and in cooperation with the Data Management Harmonization tool and the Common Operational Image (COP) tools;
Dynamic Hospital Management (CHARM): is a discrete event simulation (DES) that models the dynamic reconfiguration of hospital departments for bed capacity planning, facilitating the continuation of normal operations of the Intensive Care Unit (ICU) and epidemics;
FLEE: is an agent-based modeling code that is used to predict migratory movements and will be modified in STAMINA to include the movement of goods; Flee is used to investigate the E.Coli transmission pathway identified in the Netherlands;

The Flu and coronavirus simulator (FACS) models the transmission of SARS-CoV-2 (or other viruses) and stimulates the epidemic of SARS-CoV2 in a given region using the computational dynamics of the disease on a local scale (e.g. city or neighborhood);
BIMS: is a predictive model that simulates the transmission of West Nile virus through the use of agents that represent: the environment, mosquitoes, avian hosts and humans;

Global Epidemic and Mobility Model (GLEAM): is a stochastic tool for modeling meta-population epidemics that allows the simulation of the spatio-temporal spread of various infectious diseases globally, taking into account over 3200 subpopulations from about 230 different countries and territories worldwide;

Machine Learning-Based Early Warning System (EWS): is responsible for processing data received from various other STAMINA tools; it also produces warnings and alerts in the event of deviations from the rules or the identification of models, thus preserving not only the raw data but also the processed data;

Web and Social Media Analysis Tool (WSMA): is an online monitoring and listening tool that collects data from social media APIs based on end-user-defined search parameters;

Preparedness Pandemic Training Tool (PPT): its role is to provide the trainer (or group of trainers) with the opportunity to design a training exercise in a modern graphical user interface and to obtain the exercise scenario in a digital form, easy to edit and share;

Crisis Management Tool (CRISISHUB) has the following roles:
Response planning before a crisis;
Managing information during a crisis;
Making strategic decisions during a crisis;
(Re) allocation of resources during a crisis.
ENGAGE defines, monitors and disseminates in real time the availability of the hospital in terms of number of beds, medical services, etc.;
Common Operational Image (COP) is a web-based software solution that aims to provide different types of views by integrating external tools and data sources.

RO PILOT REQUIREMENTS AND STATUS
In the following part, we will present an example of the evolution of the covid-19 numbers of cases in Bucharest over a determined period (approximately one month).
We used Grafana to realize the analytics and show how the numbers changed by observing the values every two days.

For this graph, we used as a source the time_series_ro_counties_daily.csv file whose structure is described here. These are the steps for creating the final chart:

Create an empty time-series dashboard and select the data source.

At Path, we indicate the file.

At Fields, we indicate the columns of interest and their type.

The graph obtained at this step contains the values ​​for all cities and cannot be interpreted correctly.
Add a Filter data by Value transformation and select the data for Bucharest (iso == B)

The graph is now the desired one, the correct one, with the cases from Bucharest:

As we can see, from the start period to its end, the numbers are increasing.

The goal of the Romanian pilot is in line with the main objectives of the STAMINA [8] project and wants to implement a pandemic crisis management system, enhancing both the preparedness and the response to such a situation.
The Web and Social Media Analytics tool (WSMA) is one of the many STAMINA features that are under development. WSMA is designed to monitor and gather relevant data from social media APIs in compliance with the end-user stated search parameters. The aim of the tool is to monitor citizen sentiment and trust in institutions to facilitate the early identification of outbreaks signs.
WSMA will be used for the monitoring of social-media networks, in the first instance for Twitter and Reddit, the modularity of the tool will enable data gathering from other social-media applications as well. In addition, the users will be able to generate personalised analytics.
The WSMA tool has an important role in the STAMINA project as it helps the authorities to understand, at a more complex level, the support and trust of the citizens in the local, national and international institutions and organizations, and in the importance of the regulations and restrictions during a crisis.

Authorities can create a dashboard in this tool and add specific keywords that they want to monitor for a better understanding of the citizens and their thoughts shared on social media platforms.
A similar sentiment analysis platform was developed in the SoMeDi [14] project, which is a recruitment platform based on Natural Language Processing (NLP) that aims to explore the hidden information of the users (job applicants) regarding specific topics related to the employing company. The modularity of SoMeDi enables the project to provide relevant information about the sentiments of the citizens during this difficult period. STAMINA can build upon SoMeDi and improve the recognition of the feelings experienced by the population by crawling data from social-media APIs, processing the text and saving it into a cloud database for further analysis.
The Faculty of Political Sciences and Business Management from University Babes-Bolyai, Romania has developed a platform [9] that openly presents infographics and relevant data regarding COVID-19 and the economical impact of the pandemic on the romanian economy and some social aspects (children presence in schools, unemployment rate, time spent outside home, attendance to sport events, etc.).
Graphs.ro [15] is a website that collects and analysis public data from reliable sources, providing panels and graphics with COVID-19 informations (no. of cases, deaths, tests, vaccines, etc.) to show a comprehensive view of the pandemic situation in Romania and all the cities.
Code for Romania Association is a NGO that collaborated with the romanian government to develop a website [16] that provides informations and infographics about the pandemic situation in romania, only from official sources, to the concerned citizens and mass-media for a better understanding of the reality and to stop the phenomenon of fake-news.
User-Requirements
The requirements for the WSMA tool are categorized based on their relevance to the end-users in: Mandatory, Important and Interesting.
The mandatory ones are related to understanding public discourse about the pandemic and the responses received, identification of public health threats signals, and detecting societal distrust in pandemic measures and institutions. The important use requirements are considered: understanding the news speech and trends about the pandemics and politics, the integration with the Early Warning System so users will receive real-time alerts about the social-media trends, the detection of fake news that are distributed and the identification of the persons / organizations that are promoting them. Other important user-requirements are represented by the possibility to contextualize the relevance of the information, present the results in categories and in geographical context and to divide social-media users in different categories. As interesting user requirements it was identified the role-specific information that will be distributed based on the importance to specific users, the identification of key pathways for messages and particular keywords to be shared to social-media users, as well as for the tool to provide suggestions for the content and form of the published messages.

CONCLUSIONS AND FUTURE WORK
Various IoT solutions have been developed to enhance medical staff productivity in the healthcare sector during the years, but no complete solutions were noted. In this paper, a work in progress complex platform and its capabilities are described. The toolset will mainly include wearable and portable diagnostic devices, predictive modeling, an early warning system, and a crisis management tool. Together with the other functionalities described in the paper, STAMINA will improve relationships between the countries and increase diagnostic capability by providing tools and creating guidelines to enhance preparedness and response to worldwide crises.
IoT has rapidly entered the healthcare field because of the need for intelligent pandemic tools. The main goal is for most hospitals to access new technologies and proper knowledge to have the ability to use them. For future work, after the solution is finished, the project members will adequately test it to check for possible irregularities. The stakeholder's opinion will continuously be used for further improvements. All the facilities listed above will be implemented in easy-to-do steps to represent an excellent toolset for pandemic crisis management.

REFERENCES

Tavakoli, M., Carriere, J. and Torabi, A. (2020), Robotics, Smart wearable technologies, and autonomous intelligent systems for healthcare during the COVID-19 pandemic: an analysis of the state of the art and future vision. Adv. Intell. Syst., 2: 2000071. https://doi.org/10.1002/aisy.202000071
Greco, L., Percannella, G., Ritrovato, P., Tortorella, F. and Vento, M., 2020. Trends in IoT based solutions for health care: moving AI to the edge. Pattern Recognition Letters, [online] 135, pp.346-353.
Restauri, N. and Sheridan, A., 2020. Burnout and posttraumatic stress disorder in the coronavirus disease 2019 (COVID-19) pandemic: intersection, impact, and interventions. Journal of the American College of Radiology, 17(7), pp.921-926.
Siddiqui, M., 2020. IoMT potential impact in COVID-19: combating a pandemic with innovation. Studies in Computational Intelligence, pp.349-361.
Santamaria, A., De Rango, F., Serianni, A. and Raimondo, P., 2018. A real IoT device deployment for e-health applications under lightweight communication protocols, activity classifier and edge data filtering. Computer Communications, 128, pp.60-73.
Behera, S., et al., 2020. Biosensors in diagnosing COVID-19 and recent development. Sensors International, 1, p.100054.
Abdalla, M., Alarabi, L. and Hendawi, A., 2021. Crisis management art from the risks to the control: a review of methods and directions. Information, 12(1), p.18.
Bonavita I., Maitland E., A trust-motivated framework for assessing governments engagement with citizens on social media, in SocialSens 2021
https://econ.ubbcluj.ro/coronavirus/
Abir EL Azzaoui, Sushil Kumar Singh, Jong Hyuk Park, SNS big data analysis framework for COVID-19 outbreak prediction in smart healthy city, Sustainable Cities and Society, Volume 71, 2021, 102993, ISSN 2210-6707, https://doi.org/10.1016/j.scs.2021.102993.
Gao, Y., et al. Machine learning based early warning system enables accurate mortality risk prediction for COVID-19. Nat Commun 11, 5033 (2020). https://doi.org/10.1038/s41467-020-18684-2
Sun S, et al., Using smartphones and wearable devices to monitor behavioral changes during COVID-19, J Med Internet Res 2020;22(9):e19992, doi: 10.2196/19992
https://www.covinform.eu/consortium/
Pasat A., Suciu G., Birdici A. and Pop I., An internship campaign case study showing results of enhanced recruitment processes using NLP, in Proceedings of International Scientific Conference for eLearning and Software for Education, eLSE 2021, Vol. 2.
Vana, D., 2021. Grafice Covid-19 in Romania.. [online] Graphs.ro. Available at: https://www.graphs.ro/# [Accessed 19 August 2021].
COVID-19: Date La Zi. 2021. Date la zi. [online] Available at: https://datelazi.ro/ [Accessed 19 August 2021].

Speaker: Mr George SUCIU (BEIA Consult International)

09:40 Mediminder - Medication Management and Reminder Application

When living with a chronic disease, one's quality of life or even life itself depends significantly on the degree to which one can follow their treatments. In such cases, treatment adherence is essential as every dose counts. However, a significant number of people find it hard to follow their treatment due to the increased number of medications and administration complexity. The challenge comes with seamlessly integrating treatment management into patients' lifestyle by making it practical and accessible. The solution we propose is Mediminder, an Android application for treatments management. Designed to aid people suffering from chronic diseases, its purpose is to simplify adherence to numerous complex treatments using reminders and planning tools. We believe that complex problems have simple solutions and, with the growing number of smartphone adoptions, we see that tools in the form of mobile applications are taking over. Mediminder is one of them, designed to help more people experience the full benefits of their clinical prescriptions, lengthen their lifespan and increase the quality of their lives.

Speaker: Flavia OPREA

mediminder-roedunet.pdf

10:00 ORION – Diabetes Management Platform for Patients

The project at hand delves into the issue that any diabetic patient or their doctor has had while dealing with this disease. While diabetes has yet to be cured, the momentary resolve for it is to monitor its symptoms to subdue it. Our application accomplishes that by creating an environment where the patient can easily monitor their disease by registering their metrics into this system, while enabling their doctor to view and analyze said metrics. By providing the doctor with their patient status at any given time, the doctor can now fully tailor the treatment for the patient at hand. This project sets out to facilitate matters when dealing with the sharing of necessary data between patient and doctor, to minimize travels to the medical office just to share metrics which can be viewed through our application. The main objectives for this application have been to create an easy and user-friendly design to be pleasant to inexperienced users as well, while also being a more efficient method of registering the necessary information regarding diabetes from the patient’s viewpoint. From the doctor’s perspective we achieved a more direct means for him/her to verify its patient’s status regarding the disease.

Speaker: Flavia OPREA

orion-roedunet.pdf

10:20 Solutions for improvement the medical crisis situations management in Romania

Abstract - Are presented software and administrative-organizational solutions that support the application of national legislation, in order to increase the capacity and speed of response of public authorities in Romania with responsibilities in the management of epidemics/pandemics. In the current context, generated by the SARS-CoV2 pandemic, the proposed solutions are of high interest. The danger of the communicable diseases globally spreading is unprecedented, which requires a good organization, correlation and intervention of the authorities responsible for monitoring and maintaining public health all over the world.
The paper is an approach focused on studying the opportunity and legality of applying software or administrative-organizational solutions to improve the crisis management and response capabilities of the authorities responsible for the organization, deployment and mode of action of first responders (emergency medical technicians, paramedics, ambulance, firefighters, rescue crews and police officers).
Are briefly presented the software solutions Dynamic Hospital Ward Management model (CHARM), which aim a better allocation of hospital beds within the administrative-territorial units (counties), respectively the optimal distribution of patients from medical crisis, Crisis Management Tool (CMT ), used to assist the management of institutions/forums in charge of managing emergencies, both before they start and during crises, Web Social Media Analytics Tool (WSMA), a complementary CMT tool, with the role of increasing its efficiency . Also, the administrative-organizational solution regarding the establishment of a Reserve Structure of personnel and technical means, organized at regional level, by reference to the national legislation in force, is analyzed. The administrative-organizational solution is based on the hypothesis that the existing technical and human resources at the level of the county management unit for emergency situations (ISU) are insufficient in the case of several events, which take place simultaneously.

Speaker: Dr Tiberiu Vlad PATANCIUS (Department of Biophysics, „Stefan S. Nicolau” Institute of Virology)

#_15_Solution crisis_final_2_IVN_presentation.ppt

10:40 → 11:00 Coffee Break

11:00 → 13:00 RO-LCG 2021 "Grid, Cloud && High Performance Computing in Science" & Cloud Computing and Network Virtualisation

Chair Mihnea Dulea / Paul Gasner
RO-LCG 2021 Virtual Room A

11:00 Grid and cloud computing at INCDTIM

Grid and Cloud Computing are two different domain with the same idea, the processing and storing of data. Grid is standardized, Cloud is one step toward standardization through the new European Open Science Cloud (EOSC) project. At INCDTIM we are processing data for the last 15 years at Grid site RO-14-ITIM, but would like to add a cloud computing system at our Institute. This paper describe what we have at the Institute and what projects we are in to fulfil our long last dream of having a cloud at our location

Speaker: Dr Felix FARCAS (INCDTIM)

11:20 Testing of Grid Worker Nodes Integration in OpenStack

Testing of Grid Worker Nods Integration in OpenStack is presented.

Speaker: Dr Ciprian PINZARU (UAIC)

Testing of Grid Worker Nodes Integration in OpenStack .pdf

11:40 Big Data Performance in Private Clouds. Some Initial Findings on Apache Spark Clusters Deployed in OpenStack

In recent years Apache Spark has become one the most important Big Data platform. In-memory processing performance and the ability to connect with any major data server/source/format have been two of the main drivers of Spark’s popularity. But finding the most suitable setup for a given data processing task is challenging, depending not only on the data structure and the nature/complexity of the task, but also because of the myriad of setup parameters to be tweaked. In this paper we propose a model for assessing the processing performance of a Spark-and-Hadoop cluster, deployed on a university cloud managed with OpenStack. Randomly generated SparkSQL queries on the TPC-H benchmark schema were executed for data sets of 5GB, 10GB and 50GB, varying four data source formats and two memory settings. Predictive models built with three Machine Learning techniques (Multivariate Adaptive Regression Splines, Random Forest, and Extreme Gradient Boosting) provided encouraging results. For the given data sets, the most important predictors seem to be related with the volume of processed data and the query complexity whereas the data formats and memory settings seem less important.

Speakers: Prof. Marin FOTACHE (Al.I. Cuza University of Iasi), Mr Marius-Iulian CLUCI (Al.I.Cuza University of Iasi)

RoEduNet2021_Fotache_Cluci.pptx

12:00 Implementation of an email-based alert system for large-scale system resources

Tackling the current problems of interest for physicists that deal with various topics requires lots of computing simulations. Identifying and preventing any unusual behavior within the system resources that execute large-scale calculations is a crucial process when dealing with system administration since it can improve the run-time performance of the resources themselves and also help the physicists by obtaining the required results faster. In the present work, a simple \emph{pythonic} implementation which 1) monitors a given computing architecture (i.e., its system resources such as CPU and Memory usage), and 2) alerts a custom team of administrators via e-mail in almost real-time when certain thresholds are passed, is presented. Using existing packages written in Python, with the current implementation it is possible to send e-mails to a predefined list of clients containing detailed information about any machine running outside the "normal" parameters.

Speaker: Mr Robert POENARU (Horia Hulubei National Institute of Physics and Nuclear Engineering)

roEduNet_2021.pdf

12:20 Moving forward passwordless authentication: challenges and implementations for the private cloud

Authentication is necessary anytime we need to certify our identity. The most used type of authentication is done by using a username and a password. Historically, multi-factor authentication (MFA) came as next level of authentication. It is a great way to secure your organization, but most of the users get frustrated with the additional security feature on top of having to remember their passwords. Passwordless authentication is more convenient because the password now is removed and replaced with something you have, and it cannot be forgotten. These solutions permit users to access their workstation, application, or network by using their fingerprint, eye scan or voice recognition. Passwordless authentication nowadays is used on large corporations, and environments where authentication security is very important. As novelty, the current article is presenting a passwordless implementation solution for OpenStack private cloud. The solution can be implemented with minimal configuration changes also on other private cloud environments. A comparison between classic authentication and passwordless is made at the end of the article with the purpose of understanding the advantages of this new authentication system.

Speaker: Mr Ionel GORDIN (USV)

Moving forward passwordless authentication - Ionel Gordin - USV.pptx

12:40 Infrastructure for Capturing and Persisting Virtualization-Specific Events Triggered by In-Guest Process Executions for Behavioral-Based Analysis

Nowadays, security threats become more and more harmful. Many security solutions have been implemented along the time, the most popular being the anti-viruses. They offer proper protection against computer viruses unless these malicious programs do not run at a higher privilege level than the security solution itself. This shortcoming of conventional security solutions can be reduced using virtualization-based mechanisms, which run totally separated from the main user environment, in the same time being able to monitor events and take actions if necessary. In order to improve their performance, behavioral datasets of malicious software can be used for training a model which can then be used by the security solution. There are very few publicly known and relevant datasets from which one can build such a model, so the current paper proposes an open design for an infrastructure capable of recording and storing application behavioral events in order to train a security-oriented machine learning solution. The proposed solution consists of a hypervisor that is run on an end-user system and the necessary software that controls the activation, interception and storage of virtualization events from which one can build the relevant datasets.

Speakers: Robert VARADI (Technical University of Cluj-Napoca), Mr Gabriel RAT (Technical University of Cluj-Napoca)

11:00 → 13:00 Sensor Networking

Chair Octavian Rusu
Sensor Networking Virtual Room B

11:00 Safety and Security of Citizens in Smart Cities

Smart cities have frontline responsibility to ensure a secure and safe physical and digital ecosystem promoting cohesive and sustainable urban development for the wellbeing of human beings. In this paper, we propose to integrate advanced technological solutions in a market-oriented unified Cyber-Physical Security Management framework, aiming at raising the resilience of cities’ infrastructures, services, ICT systems, IoT, and fostering intelligence and information sharing among city’s security. The project we implement, “Smart Spaces Safety and Security for All Cities” (S4ALLCITIES), is dealing with SoS to deploy and validate its intelligent components and functionalities on actual environment, ensuring the delivery of solutions and services in line with smart cities emerging requirements, focused on: risk-based open smart spaces security management; cybersecurity shielding; and behavior tracking; real-time estimation of cyber-physical risks in multiple locations and measures activation for effective crisis management.

Speakers: Mr Ijaz HUSSAIN (Beia Consult International), Mr George IORDACHE (Beia Consult International), Mr Cristian BECEANU (Beia Consult International), Mr Robert Alfred KECS (Beia Consult International )

Presentation_Safety and Security in Cities for Smart Mobility_ id 7_RoEduNet2021.pptx

11:20 Edge computing for autonomous vehicles. A scoping review

Despite some recent notable achievements in autonomous road vehicles (e.g. the Waymo and Tesla autonomous car prototypes), the problem of pedestrian detection remains challenging and still lacks a universal solution. The vast majority of the existing solutions rely on sensors and computing equipment located on the vehicle itself, but this increases the cost and the energy consumption of the vehicle to unreasonable levels. In this context, it is worth attention the idea of relocating certain sensing and computing tasks to a network of roadside (“edge”) devices capable to communicate in real-time with a plurality of vehicles, in order to reduce the on-board equipment. This paper is a brief scoping review of the literature dedicated to this topic, aiming to: define the main concepts related to using edge computing for pedestrian detection, identify the advantages and drawbacks of this approach (especially the security issues and the means to mitigate the threats), identify the hardware needed, and outline a typical edge infrastructure of a smart intersection with pedestrian detection capabilities.

Speaker: Cristian SANDU (Universitatea "Dunarea de Jos" din Galati)

RoEduNet Edge computing. A scoping review..pptx

11:40 MEWS - an IoT and Cloud-Based avalanche detection and prediction platform

In this article we propose a Cloud-based platform for the detection and prediction of snow avalanches. The platform is based on data gathering from sensors, data collection in Cloud and using this data as an input for the developed Machine Learning algorithms. We discuss the different phases of the data workflow and present the proposed system’s functionalities. We also analyze the data validation and pre-processing operations necessary before applying the Machine Learning algorithms on the gathered data. We also present the integration of the Machine Learning capabilities with the OpenGate cloud platform.

Speakers: Dr George IORDACHE (BEIA Consult International), Dr George SUCIU (Beia Consult International)

12:00 Adaptive Scaling for Image Sensors in Embedded Security Applications

— Image sensors are widely used for multiple applications today. The security applications imply not only data acquisition, transmission, and display, but also embedded or in-the-cloud image processing algorithms as motion or intruder detection.
In this paper, we analyze the impact of image scaling on the image processing algorithms, with focus on object detection. Different methods for image scaling are presented and their impact on the algorithms are measured.
We propose an adaptive scaling mechanism, that implies the selection and dynamic adaptive configuration of one of the available hardware engines (interpolators), in order to optimize the power consumption and the bandwidth used for data transmission. The artificial intelligence processing outputs are used as a source for these configuration changes. Besides the optimal decision on enabling the appropriate re-sampling interpolator, the parameters computed (in order to configure this selected hardware block) are the quality of the scaling and the image resolution.
The hardware implementation for this mechanism is presented and the measured results are discussed.
The conclusion we drew is that an adaptive scaling mechanism in embedded security applications significantly improved the object detection algorithms and optimized the used data transfer bandwidth.

Speaker: Corneliu ZAHARIA (Transilvania University of Brasov)

12:20 An IoT Photovoltaic Sensing System

This paper follows the development of an IoT photovoltaic (PV) sensing system built around a pyranometer from Kipp and Zonen. Two Raspberry Pi act as IoT nodes connected to this system with the purpose of collecting time series values, of Global Horizontal Irradiance (GHI) from the pyranometer, and of current from a reference cell. The latter will be then calibrated with the pyranometer based on numerical methods. These time series values are sent to the Influx database cluster for storage. To ensure continuous operations for an extended period and to eliminate single points of failure, a High Availability (HA) architecture was employed. Web based Grafana acts as the monitoring solution for the collected values, stored inside the InfluxDB cluster. After obtaining a somewhat considerable dataset of irradiance-current values from the two nodes, numerical methods are then used to calibrate the reference cell. Linear regression, polynomial regression and interpolation are the methods tested. Based on observations, the numerous iterations of the models and the model performance indicators, interpolation proved to be a sound fit for calibration. The reference cell, after having learnt the values from the pyranometer will be able to swiftly and reliably determine the GHI from the current variation of its solar cell thus acting as a standalone irradiance sensing instrument. An array of such reference cells can then be calibrated to expand the IoT solar sensing network.

Speaker: Vlad VOICU (Universitatea Tehnica din Cluj-Napoca)

13:00 → 14:00 Lunch

14:00 → 16:00 Open Distance Learning && Technologies for Future Internet

Chair Lenuta Alboaie
ODL & TFI Virtual Room A

14:00 A collaborative game theory approach for determining the feasibility of a shared AS blockchain infrastructure

The paper studies the feasibility of building a shared AS blockchain infrastructure employing a collaborative game theory approach. Game theory is used due to its versatility on analyzing situation where the outcome depends on the actions of several actors (AS) and on the estimated payoff when an actor makes a decision. The games chosen are in the category of Cooperative Game type, where multiple actors cooperate and form a coalition which they join or leave based upon their payoffs / reward. PFG is the basis of the algorithm implemented to demonstrate the feasibility of integrating a blockchain solution in an AS federation and to validate the reason for this integration. A mechanism of incentives is proposed to stimulate the cooperation of multiple AS.

Speaker: Mr Rudolf Alex KOVACS (Technical University of Cluj-Napoca, Computer Science Department)

14:20 Open Distance Learning. Work from home in a pandemic context. Teaching students on the COVID 19 period.

Abstract - Working from home will become a more standardized and accepted thing; it will take a long time for people to return to the office, which will continue to focus on everything related to accessibility, remote technology and data retrieval. All those who were involved in transforming the teaching process from traditional to online, were not prepared for this and encountered a number of difficulties in this transformation. At this moment, both teachers and students had to adapt to the new way of teaching which is an alternative to consider in the future to the traditional system.

Speaker: Elena Oana CIMPIANU

14:40 Fraud Detection for Online Judge Systems

Maintaining a fair environment in the academia has
always been of high importance. Leaving ethics aside, it is a
fundamental factor for an efficient and prosperous educational
process. This process is continuously changing and evolving,
always trying to perfect itself. However, with all the benefits,
it also creates a new context for cheating and fraud, that we are
aiming to keep up with and overcome. In order to do so, we intend
to test different approaches presented in the literature, gather
their results and create an insightful set of metrics. Further,
we will use the metrics for automatically detecting fraud in the
context of Online Judge Systems. To achieve such a goal, we are
going to test, tweak some of the existing solutions and discover
links between them using a set of real submissions made by first
year students during their Data Structures practical exams.

Speaker: Serban-Ioan CIOFU

Fraud Detection for Online Judge Systems - II.pptx

15:00 Optimal Byzantine Fault Tolerance Consensus Algorithm for permissioned systems

In this paper, we propose the “Optimal Byzantine Fault Tolerance” consensus algorithm for permissioned replicated distributed systems with Byzantine fault tolerance. It seems to achieve total fairness, in the sense that it is impossible for an attacker to manipulate which of two transactions will be chosen to be first in the consensus order. It has asynchrony, no leaders, no round robin and no Proof-of-Work. A main characteristic for this algorithm is the digital signatures and broadcast usage to record network activity. The consensus participants inform everyone else about the set of accepted or rejected transactions in a time frame. As to the ordering of transactions, three mechanisms have been proposed: computer clocks, lexicographical order for message hashes and the semantic order given by the versioned data, manipulated inside transactions. Considering these traits, we can say the algorithm yields fair Byzantine agreement and a total order for all transactions, with very little communication overhead beyond the transactions themselves.

Speaker: Mrs Cristina Georgiana CALANCEA (University Alexandru Ioan Cuza Iasi, Romania)

15:20 Overview of Echo State Networks using Different Reservoirs and Activation Functions

IoT devices are growing rapidly in number and networks become more complex than ever. Therefore, network traffic prediction is a topic that needs special attention, allowing network administrators to be able to continuously adjust network parameters. This paper is focused on Echo State Networks (ESN), a quite new type of Recurrent Networks which can be used for times series prediction. Different configurations were tested, and their performance were compared to those of classical algorithms such as Seasonal Autoregressive Moving Average (SARIMA), Convolutional Neural Networks (CNN) and Long-Short Term Memory (LSTM).

Speaker: Mr Dan-Andrei MARGIN (Technical University of Cluj-Napoca)

MarginAndrei_Prezentare_RoEduNet.pptx

15:40 Short technical text classification

Recent work in machine learning has provided promising results regarding short text classification. However, in case of technical abstracts and datasets, there are some cases when unsupervised algorithms fail to reach the optimal results. This paper presents a detailed description of a text classification system for short technical texts and datasets from a single domain, together with experiments. In order to be effective, it has been built as a mix between machine learning algorithms and keywords determined by an expert’s point of view.

Speaker: Iulia-Maria Florea FLOREA (Politehnica University of Bucharest)

14:00 → 16:00 Network Management && Open Source and GNU in Education and Research

Chair Florin Bogdan Manolache
NM & OSER Virtual Room B

14:00 Automated SSL/TLS Certificate Distribution System

SSL/TLS certificates are used by more and more
network services, but their lifespan keeps decreasing. Managing
certificates on a large network is extremely time consuming,
both as manpower and as complexity. This paper presents
an automatic SSL/TLS certificate management system based
on a single certificate manager, which makes generation and
distribution of certificates efficient and traceable, while keeping
the flexibility of multiple administrators for various domains. The
system offers a web interface and a CLI, while keeping software
dependencies at a minimum. The software is used in production
on a large heterogeneous network at Carnegie Mellon University.

Speaker: Florin Bogdan MANOLACHE (Carnegie Mellon University)

14:20 Using Software-Defined Networking Technology for Delivering Software Updates to Wireless Sensor Networks

The current paper proposes a system architecture where a distributed environment of wireless sensor networks (WSNs) can selectively receive software updates when enhanced with a Software-Defined Network (SDN) control environment. In the context of IoT, the current work aims to facilitate the deployment of software updates on WSNs in an automated fashion and also proposes the usage of this architecture in large scale business networks, where different areas of the same network may change their purpose at different times during the network lifecycle. A central component oversees the coordination of each WSN, but each WSN is responsible for retrieving its updates from the storage location. For security, the software updates are stored in a remote peer-2-peer storage location. A simulation environment is also presented which uses Mininet-Wifi as a WSN emulator with Pox controller as SDN enabler and uses the IPFS network as a remote peer-2-peer storage. The Pox controller’s host_tracker module is enhanced with features to retrieve the updates from the IPFS network and to deliver the retrieved information to each station on the network. The simulations show that information can be delivered with relatively small network overhead and changes to the Pox controller, making this a viable solution for delivering updates to WSNs.

Speaker: Sorin BUZURA (Technical University of Cluj-Napoca)

RoEduNet_2021.pptx

14:40 Improving Upon Photographic Steganography

Extensive research encompassing steganography has been carried out in the past couple of decades and has spurred numerous studies. Research regarding steganography and steganalysis has recently shifted, and impressive progress has been carried out with the aid of the emerging popularity of deep convolutional neural networks. This paper addresses a new branch of steganography, namely photographic steganography, which intends to approach hidden message recovery from a steganographic image captured on a consumer-grade device. Challenges faced by this technique require the modeling of image perturbations to facilitate accurate message retrieval, and several papers have thus far managed to obtain impressive results in various image capturing conditions and in bit recovery error rates. However, several limitations persist, such as short hidden message lengths and restrictions imposed upon encoded images. This paper addresses the shortcomings, and novel techniques to improve the current state-of-the-art will be proposed to offer a more complete, accessible, and performant solution.

Speaker: Giorgiana VLĂSCEANU (University Politehnica of Bucharest)

15:00 Towards the implementation of FAIR principles on an earthquake analysis platform

FAIR principles of scientific data represent a relatively new direction in research that is best described as an initiative to encourage and help researches in making their data more available and easier to reproduce and reuse. Most importantly, the principles emphasize machine actionability, meaning the data requires strict guidelines so that automatic systems can read, retrieve and use the data with minimal human intervention. FAIR stands for Findable, Accessible, Interoperable, and Reusable, defined as below:

• Findable : data must be easy to find for humans and machines.
  Databases must be described with an abundance of metadata and clear,
  explicit identifiers so that they are machine readable. The storage
  system must be capable of discovering the data by intuitive query.
• Accessible : after finding it, the users must know how to access it,
  possibly through authentication and authorization, using a
  standardized communications protocol. This protocol must be open,
  free and universally implemented.
• Interoperable : metadata and databases need to respect the same
  format, so that they can be integrated with each other and that you
  can use the same analysis, storage or processing workflows,
  interchangeably.
• Reusable : data must be well-described and commented so that they can
  pe reproduced or combined. They will be described by the same
  relevant attributes and will be released with a clear and accessible
  data usage license. Finally, the details of their collection must
  meet the standards of the scientific community which they are meant
  for.

In our work on earthquake analysis the first step was that of collecting, parsing, curating and storing the seismic databases available online. Because of the small differences in the structure of the publicly available earthquake databases, our job was to select only the relevant information present in all databases. To this end we developed a parsing code which downloaded the databases, retrieved the data, and created new databases on a new format that allowed total interoperability between them.

The next step was to develop the analysis tools. Here, through the theory of complex networks, we elaborated codes which call the databases and use the information to create a seismic network by splitting the seismic region into small 3D cubes, and placing each earthquake in their respective cube, based on its epicenter geographical position. Building the network chronologically, each subsequent earthquake represents a link in the network, whereas the cubes represent the nodes. For these seismic networks we developed a series of codes which analyse the degree of connectivity, the structure of the motifs, and allow for automatic visualizations using open source third party software.

Summing up, we report a new analysis framework for earthquakes in which data gathered from public sources is structured in a database format that allows automatic processing through a series of codes developed in house. This process represents a “FAIRification” of the available seismic data, as the databases that we created preserve only a fraction of the initial information, but allow Interoperability and Reusability of these databases, in addition to the Findability and Accessibility that was already provided for by the original seismic catalogues. This opens new areas of research associated to FAIR and Open Science, the research on the IT and Computing infrastructure supporting these principle, research on designing faster and reliable networks for supporting access and the design and implementation of new cybersecurity standards and controls that could support the users.

Speaker: Gabriel PANA (Faculty of Physics, University of Bucharest)

15:20 Automatic Integration of D Code With the Linux Kernel

The Linux kernel is implemented in C, an unsafe programming language, which puts the burden of memory management, type and bounds checking, and error handling in the hands of the developer. Hundreds of buffer overflow bugs have compromised Linux systems over the years, leading to endless layers of mitigations applied on top of C.

In contrast, the D programming language offers automated memory safety checks and modern features such as OOP, templates and functional style constructs. In addition, interoperability with C is supported out of the box. However, to integrate a D module with the Linux kernel it is required that the needed C header files are translated to D header files. This is a tedious, time consuming, manual task. Although a tool to automate this process exists, called DPP, it does not work with the complicated, sometimes convoluted, kernel code.

In this paper, we improve DPP with the ability to translate any Linux kernel C header to D. Our work enables the development and integration of D code inside the Linux kernel, thus facilitating a method of making the kernel memory safe.

Speaker: Constantin Eduard STANILOIU (University POLITEHNICA of Bucharest)

15:40 On How To Combine Image Segmentation Algorithms Using Entropy

Image segmentation is one of the most frequently used computer vision techniques. Whether we talk about medical imaging or autonomous driving, image segmentation algorithms are required to obtain the desired result. Therefore, a variety of algorithms have been implemented so far. Being based on different approaches, each of these algorithms has its own advantages and disadvantages. No algorithm can perform the same regardless of input data, with algorithms yielding better or worse results depending on the characteristics of the image. Some may accurately preserve the borders between large regions while clustering small details together (under-segmentation), while others can correctly delimit details while at the same time splitting large regions in multiple clusters (over-segmentation). Moreover, some algorithms might have a natural tendency in over-segmentation or under-segmentation, independent of input. This paper proposes a voting method which combines the results of some notable segmentation algorithms. The aim of this method is to limit the downsides of the used algorithms and to obtain, if possible, more accurate results. The results show that, in most cases, the proposed method offers both improvements in the quality of the provided output and more overall confidence in its usage.

Speaker: Mrs Giorgiana Violeta VLĂSCEANU (University Politehnica of Bucharest)

16:00 → 18:00 Technologies for Future Internet

Chair Florin Manolache
TFI Virtual Room A

16:00 Approaching traffic congestion with Double Deep Q-Networks

One of the recurring problems of everyday life in big urban areas is the traffic congestion. Nowadays advance in the technologies powering the Internet of Vehicles together with state-of-the-art artificial intelligence algorithms offer the means to improve traffic flow with little changes to the existing infrastructure. This paper proposes a reinforcement learning based solution to traffic light scheduling with a case study on four notoriously congested traffic areas from Bucharest, Romania. The selected areas were imported from OpenStreetMap (OSM) files and modeled using SUMO (Simulation of Urban Mobility). We propose a C-ITS based protocol for information exchange between cars and infrastructure (Vehicle-To-Infrastructure - V2I communications) and we generate data for training a semi-centralized reinforcement learning agent that is able to change the traffic lights from its designated area. We compare our solution with fixed time slice traffic light scheduling (considered baseline performance) and we report the results. This comparative analysis shows that the proposed approach outperformed the baseline.

Speaker: Mirabela MEDVEI (Politehnica University of Bucharest)

16:20 Deep Learning for Forecasting the Energy Consumption in Public Buildings

In this paper we propose a deep learning based method to forecast the energy consumption in public buildings, based on past measurements. The method integrates two neural networks, namely a Feed-Forward Neural Network and a Long-
Short Term Memory Network. Our approach consists of three main steps: data processing, training and validation, and finally the forecasting step. We validated the method on a data set consisting of measurements taken every half an hour from the main building of the National Archives of the United Kingdom, in Kew.

Speakers: Dr Viorica CHIFU (Technical University of Cluj-Napoca), Dr Cristina Bianca POP (Technical University of Cluj-Napoca), Dr Emil Stefan CHIFU (Technical University of Cluj-Napoca)

RoeduNet-Deep-Learning.pdf

16:40 Benchmarking privacy in text classification

In most Machine Learning models, the data used for training or testing is public, available to anyone who wishes to see it. New research has improved these models, by adding privacy and distributing the processing load on multiple workers in the cloud. The aim of the paper is to perform an analysis between a classical approach (in which we have access to all the data) and one in which the privacy is preserved (Federated Learning) to explore the cases when a private model can be suitable in real-life scenarios.

Speaker: Iulia FLOREA (Politehnica University of Bucharest)

17:00 Forecasting the Short-Term Energy Consumption Using Random Forests and Gradient Boosting

This paper comparatively analyzes the performance of two machine learning algorithms (i.e. Random Forests and Gradient Boosting) in the field of forecasting the energy consumption based on historical data. The two algorithms are applied
in order to forecast the energy consumption individually, and then combined together by using a Weighted Average Ensemble Method. The comparison among the achieved experimental results proves that the Weighted Average Ensemble Method
provides more accurate results than each of the two algorithms applied alone.

Speakers: Dr Emil Stefan CHIFU (Technical University of Cluj-Napoca), Dr Viorica Rozina CHIFU (Technical University of Cluj-Napoca), Dr Cristina Bianca POP (Technical University of Cluj-Napoca)

RoeduCorina2021.pdf

17:20 Swarm Communication using Self Sovereign Identities

Swarm communication represents a communication paradigm that offers support in building and running executable choreographies that can be used to model business workflows. In the current paper we present how Self Sovereign Identities (SSIs) can be used to improve the Swarm communication. The usage of the SSIs allows us to have a decentralized way to identify entities involved in Swarm choreographies, improving the security and privacy of the choreographies and their interactions.

Speaker: Nicu-Cosmin URSACHE (Faculty of Computer Science, Alexandru Ioan Cuza University)

16:00 → 18:00 Network Security && Pervasive Systems and Computing

Chair Razvan Rughinis
NS & PSC Virtual Room B

16:00 Extending Client-Server API Support for Memory Safe Programming Languages

Google web applications have become an integral component in the day to day life of both organizations and individuals alike. These may be accessed through the
graphical user interface (GUI) or through the application programming interface (API). The latter is primarily used by programmers to integrate such services into
their applications.

Most of the languages used to implement such applications are designed with performance in mind, often neglecting security. However, security has become a major concern for such systems thus increasing demand for memory safe languages. Unfortunately, languages such as D and Rust, known to be memory safe, are lacking support for Google services.

To that end, we develop a methodology of integrating Google services with safe programming languages. We show that the D programming language can easily and successfully integrate such services bringing a boost in security and productivity.

Speaker: Constantin Eduard STANILOIU (University POLITEHNICA of Bucharest)

16:20 Security Audit for the D Programming Language

Memory corruption has been, traditionally, the number one cause for software vulnerabilities. As a consequence, programming languages that offer automated, compile time memory safety checks have been developed, such as D and Rust. However, since programming languages are pieces of software, they also may suffer
from vulnerabilities that may be exploited to bypass the memory safety checking algorithm.

In this paper, we perform a security audit of the D programming language. Our findings uncover security holes in the D safety checking system. We show
that it is possible to escape expired stack pointers which can be used to ultimately execute arbitrary code. In addition, we discuss and implement potential fixes to the discovered issues.

Speaker: Razvan NITU (University POLITEHNICA of Bucharest)

16:40 Cloud based mobile application security enforcement using device attestation API

Abstract—Today the mobile devices are more and more
present in our lives, and the mobile applications market has
experienced a sharp growth. Most of these applications are
made to make our daily lives easier, and for this a large part of
them consume various web services. Given this transition, from
desktop and web applications to mobile applications, many
critical services have begun to expose their APIs for use by such
application clients. But unfortunately, this transition has paved
the way for new vulnerabilities, vulnerabilities used to compress
cloud services. In this article is analyzed which are the main
security problems and how they can be solved using the
attestation services, meaning those services that indicate that the
device running the application and the client application are
genius.

Speakers: CRISTIAN CONTASEL (University Politehnica of Bucharest), Dr Dumitru Cristian TRANCĂ (University POLITEHNICA of Bucharest, Computer Science and Engineering Department)

17:00 Building an Interface for the D Compiler Library

The D programming language has been at the forefront of the memory safe programming languages scene. However, its adoption has been hindered by the scarce availability of 3rd party tools that aid software development such as: linters, code analyzers, integrated development envirnments (IDE) etc. The fundamental reason for this absence is the underdeveloped compiler library, that does not offer a flexible, easy to use interface and misses some of the important features such as symbol resolution and scope retrieval.

In this work, we aim at improving the D compiler library by defining a proper interface and adding the currently missing features. To understand what are the needs of a well rounded compiler interface, we analyze existing 3rd party tools and extract a series of common needs. Further, we implement these interfaces directly into the source code of the reference D compiler, with no performance loss. Finally, we upgrade existing tools to use the newly defined compiler interface. By doing so, we unify the different implementations of the same logic. This has the advantage that once the compiler is upgraded to a newer version, all the tools that leverage the compiler library will no longer need to be updated.

Speaker: Razvan NITU (University POLITEHNICA of Bucharest)

17:20 Indoor Positioning for Low-Cost IoT Devices

Precise positioning is traditionally accomplished by ground- or satellite-based navigation systems, but these are not usually available indoors. Indoor positioning systems have been demonstrated that rely on hardware normally used for wireless communication. These are either imprecise or require the use of costly high-performance hardware. We demonstrate the feasibility of decimeter-level precision for indoor positioning with the use of cost-optimized hardware specific to IoT nodes.

Speaker: Dr Dumitru Cristian TRANCĂ (University POLITEHNICA of Bucharest, Computer Science and Engineering Department)

17:40 Architecting a scalable e-election system using Blockchain technologies

The ability to vote in state elections represents a basic right and expressing the vote is a civic obligation. However, the pandemics of COVID-19 showed us how such a simple process for a citizen may have multiple implications and impose indirect barriers in expressing the electoral options. The risk of getting in contact with the virus, social distancing, or the reduced number of electors in the voting centre made the classical voting method infeasible in times of crisis. Therefore, a new opportunity appears in the market, anticipated since 2016 by the Eurobarometer - the need to provide a way to vote over the Internet. Currently, in the European Union, only Estonia can provide citizens with such services and only nine states allow voting through the postal services. To overcome the current situation, we propose an architecture of a highly scalable e-election system implemented over the blockchain technology. The aim of the project is to provide a general viable solution and to reduce the time to market, as a quick response to the current situation. The latter aspect is ensured by Hyperledger Fabric, a permissioned, highly customizable, and integration-focused blockchain implementation with consistent impact in key and complex industries like supply-chain, finance, or healthcare. The current paper elaborates a concept architecture and a corresponding design view with Hyperledger Fabric. Furthermore, it brings a qualitative solution evaluation covering security aspects (e.g., resilience to Sybil Attack) and performance measurements (block sizing, state database performance). In contrast to other existing platforms, we match together blockchain with a simple but tamper-proof, well distributed, and highly isolated architectural view.

Speaker: Ioan-Mihail STAN (University Politehnica Bucharest)

Architecting a scalable e-election system using Blockchain.pdf

18:00 → 18:20 Closing Plenary Session

Chair Octavian Rusu
Closing - Plenary Virtual Room