2023 RoEduNet Conference: Networking in Education and Research

21 Sept 2023, 09:00 → 22 Sept 2023, 17:00 Europe/Bucharest

Aula Mihai I (University of Craiova)

22nd 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 22nd edition of Agency ARNIEC/RoEduNet's (Romanian Education Network) annual Conference is organised with the help of "University of Craiova",  "Alexandru Ioan Cuza" University of Iasi, Carnegie Mellon University, Pittsburgh, US, 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.  

You are welcome to join in. Please register as a user on the conference website to receive updates on the 22nd RoEduNet Conference.  

The conference will be organized as an in-person event with help from our partners and sponsors.

Thank you, and hope to see you soon at the 22nd Conference of RoEduNet !!!

CfP 20230619.pdf

Conference Agenda.pdf

Thursday 21 September

09:00 → 13:00 Opening Plenary Session

Opening Plenary Session

Convener: Dan MANCAȘ (Agency ARNIEC/RoEduNet, University of Craiova)

09:00 Registarion

10:00 Opening / Welcome

Welcome message from conference hosts

Speakers: Cezar SPINU (University of Craiova), Gheorghe DINU (Agency ARNIEC / RoEduNet)

10:20 RoEduNet in RoNaQCI

Speaker: Mihai CARABAȘ (Agency ARNIEC/RoEduNet, University Politehnica of Bucharest)

10:40 Wi-Fi Network Monitoring with GÉANT WiFiMon

Speaker: Nikos KOSTOPOULOS (National Technical University in Athens)

11:00 NMaaS by GÉANT

Speaker: Jovana Vuleta-Radoičić (Institute Mihajlo Pupin, Belgrade)

NMaaS - RoEduNet.pdf

11:20 Coffee Break

11:30 AI/ML@Cisco

Cisco partner

Speaker: George BOULESCU (CISCO)

12:00 Innovation and Expertise in IT&C Solutions for Educational Environment

Speaker: Petre POPESCU (Datanet System)

12:30 Today’s learners. Tomorrow’s innovators

Speaker: Andrei UHR (Dell Technologies)

13:00 → 14:00 Lunch

14:00 → 18:20 Session A

Session A

Convener: Răzvan Victor RUGHINIȘ (University Politehnica of Bucharest)

14:00 Optimizing large data transfers for the ALICE experiment in Run 3

As the physics scopes of High Energy Physics experiments grow, so do their
storage needs increase. Coming with Run 3 for ALICE and LHCb, and Run 4 for
CMS and ATLAS, the rate and volume of data recorded through the detectors has increased
or will be increased to such a degree that old data management strategies
will not make due. Since the beginning of the WLCG each experiment has had its
own way of managing data, and this management software has evolved based on
the needs of each experiment. Some software being focused on decentralized
management based on datasets and data collection, and other software, such as
the one in ALICE, being focused on a structured, centralized management
infrastructure.

In this article we will present the way in which the ALICE experiment has
planned and managed data movement from the raw data processing farm to the
different storage mediums, focusing on optimizing and automating data
transfers between different storage systems.

Speaker: Sergiu Weisz (University Politehnica of Bucharest)

14:20 Enterprise Data Collection and Cross-Referencing System

Enterprise environments involve a large set of servers and databases
containing fluid and sometime inconsistent information regarding people,
infrastructure, inventory, and organizational structure.
IT support must have access and use all this information
to properly offer the right services and preemptive actions.
This paper presents a flexible and secure data collection
system which offers quick access to all the available information,
and can be used efficiently for IT and all levels of enterprise
management.
Various components of the infrastructure can easily push the relevant
information into the system, and the users have layered access
to see, edit, filter, and cross-reference the data of interest for their
function.
Triggers for automatic actions when certain conditions are fulfilled
can be configured, for signaling the need for human action.
The system was built and tested at Carnegie Mellon University.

Speaker: Florin Bogdan MANOLACHE (Carnegie Mellon University)

14:40 Automated Infrastructure Provisioning

My diploma project is focused on developing an automated infrastructure provisioning that simplifies the process of setting up and managing a network of virtual machines. The infrastructure will utilize virtualization technologies to create a cluster of machines that can be quickly and easily configured to meet the needs of the user. The system will also provide tools for managing the machines and their associated resources.

Speaker: Prof. Mihai CARABAS

15:00 Electronic steering wheel develompent for a Formula Student race car

As technology and the pace of change continue to
advance, the academy faces the challenge of preparing students
for future employment markets characterized by change and
uncertainty. To succeed in this market, practitioners need to
develop a broad range of skills in addition to more refined
practices within traditional disciplines. In this paper, we discuss
the benefits achieved through design and development of a
steering wheel among students participating in the Formula
Student project, an international competition where teams
design and build a racecar while overcoming challenges faced
along the way. We analyze the framework, embedded
component arrangement, methods and processes used by the
Bucharest UPBDrive racing team, along with the outcomes and
critical lessons learned for the next competition.

Speaker: Alexandru Vaduva (UPB)

15:20 Coffee Break

15:40 Overcoming Challenges in Migrating Modular Monolith from On-Premise to AWS Cloud

Cloud computing has gained popularity for efficient data storage, processing, and application access. This study focuses on migrating Hermit Portal's modular monolithic web application from on-premises to AWS cloud. The aim is to identify migration patterns, optimize costs, and consider project management constraints (e.g. time, cost, and performance).
Considering the Cloud Computing models perspective, the study analyzed Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Serverless models, with a focus on Amazon Web Services (AWS). Initial “lift and shift” using IaaS is recommended, followed by a gradual adoption of PaaS components like AWS Elastic Beanstalk or AWS Lambda for simplified management and reduced infrastructure responsibilities. Additionally, we explored various cost optimization strategies specific to AWS. Utilizing Reserved Instances, optimizing Windows Server and SQL Server licensing, considering PostgreSQL as an alternative to SQL Server, and leveraging Linux-based apps and web servers on AWS, were found to be effective in reducing costs.
Also, we realized during the research that conducting a comprehensive analysis within the AWS ecosystem and understanding AWS-specific features before migration is crucial. This avoids unexpected cost increases compared with on-premises hosting and accounts for refactoring and acquiring AWS-specific expertise.
In conclusion, based on our research, a phased approach with thorough analysis and AWS-Specific strategies is the key to successfully and cost-effectively migrating monolith applications to AWS Cloud. Choosing the right cloud computing model and implementing cost optimization techniques ensure a smooth transition and maximize the benefits of AWS

Keywords - Cloud Migration; Modular monolith; Lift and Shift; Clean Architecture; Optimizing costs; Optimizing resources; Database migration; AWS Pricing

Speaker: Florin Olariu (line 2: Faculty of Computer Science line 3: Alexandru Ioan Cuza University line 4: Strada General Henri Mathias Berthelot Nr. 16, Iași 700259, România line 5: olariu@gmail.com)

16:00 A Study of GPU Memory Vulnerabilities

Abstract—GPUs are a fundamental component of most high
computing environments available nowadays, being a general
solution to various purposes that range from scientific solutions
designed as software applications to clusters that hold and process
sensitive data. In this paper we focus on Buffer Overflow attacks
and memory vulnerability analysis with the intention to prove
that given special circumstances the stack and heap zones are
susceptible to memory overflow due to a lack of protection
mechanisms.
Index Terms—GPU, GPGPU, Buffer Overflow, CUDA, Stack
Overflow, Heap overflow, Memory Leaks

Speaker: Ruxandra Chiroiu (Computer Science Department University Politehnica of Bucharest)

16:20 QGP-VPN: QKD enhanced VPN solution for general-purpose encrypted communications

Quantum-resistant encrypted communications is of
particular importance in the context of recent advancements in
quantum computing. Existing technologies such as Quantum Key
Distribution achieve this, but are not suitable for general-purpose
communication due to limited key rates. In this work, we propose
a VPN solution enhanced by QKD for quantum-resistant en-
cryption, we implement it using IDQ Cerberis XG QKD devices,
and we test its general-purpose capabilities by performing an
encrypted video-conference between two endpoints.

Speaker: Alin-Bogdan POPA (University POLITEHNICA of Bucharest)

16:40 A Survey on Security Issues in Cloud Systems

—"The cloud" points out to servers, the databases and software that run on those servers. The Internet has always been constructed of infrastructure, servers and clients. Clients send their requests to servers, and servers reply, but cloud computing is distinct in the form that cloud servers are not just replying to requests with on-demand processing but they are saving data and running programs instead of the client. Cloud computing is a technology that is crossing a great expansion today. In cloud computing you can dynamically expand resources without knowledge of a new infrastructure, without developing new software or preparing new staff and to access such great technology the only things that are asked for are an Internet connection and an Internet browser. The goal of this survey is to present the different cloud security threats and recognize the proper security mechanism used to reduce them.

Speakers: Mr Diyar Jamal Hamad (Computer Science, Politehnica University of Bucharest, Romania. IT department, Soran Technical College, Erbil Polytechnic University, Iraq), Ms Khirota Gorgees Yalda (Computer Science, Politehnica University of Bucharest, Romania. Soran Technical College, Erbil Polytechnic University, Iraq)

17:00 An Email Classification Framework for Phishing Detection in Virtualized Network Environments

Phishing attacks pose a significant threat to network security, and effective detection and mitigation techniques are crucial to safeguarding sensitive information. This paper presents a novel email classification framework for phishing detection in virtualized network environments with a specific emphasis on distributed firewall architectures. The framework leverages the power of the Naive Bayes classifier to analyze the preprocessed text of emails and accurately classify them as either phishing or legitimate. By integrating the framework into a distributed firewall setup, the email traffic can be dynamically controlled and filtered at multiple network points, enhancing the overall security posture. The script accompanying this paper demonstrates the practical implementation of the framework and provides an example of its usage in a distributed firewall setup. Experimental results showcase the framework's effectiveness in accurately classifying emails, thereby enabling the deployment of appropriate firewall rules to block phishing emails from reaching their intended recipients. The framework's modular design allows for easy integration into existing virtualized network architectures, providing an additional layer of security against email-based threats. Overall, this research contributes to the advancement of distributed firewall technologies and strengthens the email security infrastructure in virtualized network environments.

Speaker: Mr Andrei-Daniel Tudosi (USV)

An Email Classification Framework for Phishing Detection in Virtualized Network Environments.pptx

14:00 → 18:20 Session B

Session B

Convener: Emil-Ioan CEBUC (Agency ARNIEC/RoEduNet, Technical University of Cluj-Napoca)

14:00 Practical Approach to Design and Implement a P2P and E2EE Instant Messaging System

As the internet became more and more globally accessible, so did messaging systems become more popular. Starting with mailing lists, IRC (Internet Relay Chat) channels and then evolving into instant messaging applications, the privacy aspect of the communication became an ever increasing concern when it comes to such applications. Many related applications and research studies tackle the problem of privacy with different approaches. The current article discusses the challenges in designing an open source, P2P (peer-to-peer) and E2EE (end-to-end encrypted) instant messaging system and offers technical details on how to implement and design a working proof of concept that allows message exchanges in real-time. The implemented proof of concept solution offers private conversations, while requiring no central servers or data silos, all the data is user hosted.

Speakers: Adrian PECULEA (Technical University of Cluj-Napoca), Emil-Ioan CEBUC (Agency ARNIEC/RoEduNet, Technical University of Cluj-Napoca)

14:20 Architecture, Design and Implementation of a Modular Solution Prototype for Flood and Fire Risk Mitigation in Wetlands

Our paper presents the architecture, design, and some implementation details of a system aimed at using two novel fire and flood sensors based on miniature sensing technology, processing the real-time acquired data using business intelligence tools for detection of sudden increases in temperature and water levels, and generating visual alerts to wetland administering authorities for better crisis assessment and early response to wetland threats. Besides customizing and connecting the sensors, a communication infrastructure was created, including connectivity with commercially available sensors, the database, the user (through a graphical interface), and modules to process the sensors data and generate alarms. We also developed and integrated a simulator for real-time sensor data. This distributed solution was virtualized in Docker with the aim that potential users could download our work and, in a guided way, be able to redo it. We look forward to developing more modules for monitoring and eco-actions in the Danube Delta, targeting other threats like poaching, fish and bird population dangers, vegetation issues, pollution sources, etc., while also working with wetland administration authorities for better crisis assessment and early response to wetland threats.

Speaker: Eduard-Cristian Popovici (University Politehnica of Bucharest)

14:40 Architecture for Inspecting Bluetooth Traffic in Software-Defined Networks

Software-Defined Networking (SDN) gained popularity in recent years for network management, network function virtualization (NFV) and its integration in wireless networks. The current works proposes a traffic anomaly detection solution for Bluetooth (BT) networks that is based on an SDN architecture. It is focused on identifying potential security threats. The SDN context provides a fully configurable and dynamic environment to monitor the traffic. The solution relies on Layer 2 information found in the BT packets, more specifically, information regarding transmission signal strength, noise and packet frequency. Using an Ubertooth One monitoring device, packets are intercepted in the SDN data plane and are forwarded to the SDN control plane for traffic analysis. Practical experiments were made using several BT devices placed at different transmission distances. The performance of the solution is analyzed and compared to other results from related studies. The experimental results offer pertinent insight, as on average, the threats can be identified in under 10 milliseconds.

Speakers: Adrian PECULEA (Technical University of Cluj-Napoca), Emil-Ioan CEBUC (Agency ARNIEC/RoEduNet, Technical University of Cluj-Napoca)

15:00 Smart Grid advancements based on the HummingBoard solution

Smart grids represent a highly sophisticated technological innovation that enables the intelligent integration of energy generators, consumers, and storage systems. This integration facilitates the provision of electricity that is both efficient and of high quality, while minimizing losses and ensuring the security of power supply. Despite its potential benefits, this technology is susceptible to security breaches, privacy issues, and vulnerabilities. The TESTBED2 initiative is a comprehensive research endeavor that spans various fields of study throughout Europe. The main aim of this study is to develop novel methodologies to enhance the scalability of smart grid services through the utilization of emerging technologies, such as blockchain. The present study expounds upon the fundamental framework of the proposed system, which employs blockchain technology as a means of certification authentication. This action was carried out within the framework of the research paper. In addition to the system description, we present the HummingBoard implementation as well as smart grid authorization policies in this paper.

Speaker: Cosmina Stalidi (Beia Consult International)

15:20 Coffee Break

15:40 Limit order for Decentralized Exchanges on Blockchain

Decentralized Exchanges allow users to trade assets using Automated Market Makers (AMM) with a smart contract. This enhances the traditional way of trading because there is no schedule (DEX is open 24/7), nor a counter-party required for an exchange.
This paper addresses the challenge of implementing Limit Order functionality on Decentralized Exchanges (DEXs). This allows users to create an order at a certain value. We developed this solution for the MultiversX blockchain, which right now allows users to trade only at the current price. We proposes a solution in the form of a decentralized orderbook. In this model, limit orders match each other (when buy and sell prices align) or serve against an Automated Market Maker (AMM) upon meeting predefined trigger conditions. Importantly, all orders are stored and settled on-chain. Our model incorporates a Rust Smart Contract for order management and a Python-based Keeper Bot for real-time market monitoring. Moreover, we introduce special economic incentives to mimic a Centralised Limit Order Book (CLOB),
offering greater flexibility and control to traders.
Our research concludes that introducing a decentralized orderbook can significantly boost trading efficiency in DEXs, thus underscoring the viability and benefits of our proposed solution.

Speaker: Mr Răzvan-Andrei Tănase (University POLITEHNICA of Bucharest)

16:00 Removing individuals from video streams through facial recognition

At the current moment, with the accelerated development of video streaming systems due to the pandemic context, more and more people have been forced to transition to the online environment in order to continue their previous routines. To minimize the feeling of being online, most meetings (e.g., sessions, courses, labs) were conducted with the camera turned on. To maintain a level of privacy, streaming platforms provide virtual backgrounds, but they only cover static objects, not people. In this article, we propose a solution where only the reference person remains in the frame while the other concurrent participants are removed from the frame, while still maintaining the virtual background. Through the proposed solution, which integrates a series of image processing algorithms that support the process of detecting people in the video stream and identifying the main speaker, we bring an added level of privacy for participants in video meetings.
Keywords: facial recognition, removing individuals from videostream

Speakers: Ecaterina Hrib (Alexandru Ioan Cuza University), Lenuta Alboaie (Alexandru Ioan Cuza University of Iasi)

16:20 Data Acquisition System for Forecasting Applications of Photovoltaic Power Production

Photovoltaic power production is dependent on atmospheric factors that change rapidly. Therefore a weather monitoring system is required to better observer modifications that can alter the power production of a PV power plant. This paper proposes a low-cost data acquisition system constructed with the help of a Raspberry Pi. Additional sensors included are BME280 for temperature, humidity and pressure monitoring, a fisheye camera for all-sky imaging and a low-cost SDR system user for satellite imagery acquisition. The presented study is used only for the data acquisition part of the forecasting methodology.

Speaker: Vlad Voicu (Technical University of Cluj-Napoca)

16:40 Real-Time Notifier for Blockchain Events

Blockchain technology gained much traction in the last few years. These decentralized databases offer security, immutability, and scalability across various applications. Decentralized applications generate vast amounts of data, known as events, that are recorded on the blockchain and are public to anyone. Some people may see opportunities for financial gains in these events and would like to know when they occur. This paper proposes a solution to process and deliver those events as real-time alerts to the users. It uses existing technologies such as message queues, multi-threading, and asynchronous processing and integrates them into a scalable architecture. The results we achieved in this paper show that for an evenly distributed network traffic, which does not entirely consists of transaction bursts, the proposed solution offers reliability, efficiency, and a suitable delivery time to those wishing to integrate it into their projects. With time, this solution, or improved architectures, may form the basis of the following big-data architectures for processing blockchain events.

Speaker: Mr Cristian-Tănase PARIS (University POLITEHNICA of Bucharest)

17:00 Awareness on Spear Phishing Attacks using Novel TLD names in Research and Academic Field

This paper conducts a study upon possibilities and
dangers of fraudulent attacks that can be performed upon
academic and research platforms, with the usage of the novel
top level domain names issued by Google Registry: .dad, .phd,
.prof, .esq, .foo, .zip, .mov and .nexus in May 2023. In the
following sections, we will focus on attempts using .zip TLD. The
present contains a bidirectional perspective simulated attack -
one in which the academic institution represents the target and
another in which the said institution represents the impersonated
entity, as well as analysis upon platform decisions of main public
communication channels and manners of protection against data
fraud attacks and attempts in academic and research scenario.

Speaker: Cristiana Constantinescu (Scoala Doctorala “Constantin Belea”, Universitatea din Craiova)

19:30 → 21:30 Welcome Cocktail

Friday 22 September

09:00 → 12:20 Session C

Session C

Convener: Mihai CARABAȘ (Agency ARNIEC/RoEduNet, University Politehnica of Bucharest)

09:00 A Case Study on Extending the Use of Cloud-Based Services within two Romanian Higher-Education Institutions

The need for digitalization of education services is obvious, considering the natural and social challenges of humankind in recent years. To face these challenges, actions are needed to improve the efficiency (or automation) of education services. This is relevant given the important role of education in society. Through this article, a technical investigation is carried out on the need for digitalization in education, and on the opportunities and advantages that arise through it. A cloud-oriented middleware architecture is proposed regarding the digitalization of certain services in education, especially those related to e-secretariat (use of digital signature and streamlining of document flows), processing of students' information (providing support for learning and teaching) and voluntarily certifying content posted online (as a countermeasure against AI-based plagiarism and fake news propagation).

Speaker: Marius Marian (University of Craiova)

09:20 A Conceptual Model for a Cloud-Based Digital Signing Service

Today, organizations around the world are using less and less paper documents that have handwritten signatures accompanied by stamps to prove their authenticity and are increasingly turning to digital signatures. Initially, the digital signature could be applied to documents by means of a hardware token, containing the cryptographic material (the private key and the associated digital certificate) together with the software for applying the digital signature. Lately, to further ease the process of applying the digital signature, the remote signature is gaining momentum. This paper outlines the advantages of using digital signature as a cloud-based service, and the path taken to achieve an experimental prototype for the digital signature service in the cloud.

Speakers: Marius Marian (University of Craiova), Mr Adelin Cusman (CERT DIGITAL), Dr Ilie Borcoși (Constantin Brancusi University)

09:40 Improving Network Security Using DD-WRT as a Solution for SOHO Routers

Improving Network Security Using DD-WRT as a Solution for SOHO Routers

Timofte Edi Marian
”Ștefan cel Mare” University
Suceava, Romania
edi.timofte@usm.ro

Balan Doru
”Ștefan cel Mare” University
Suceava, Romania
dorub@usv.ro

Abstract—Securing the IT infrastructure in today's digital age against cyber-attacks has become vital, one of the most used robust solutions for cyber security being a DD-WRT solution. This article focuses on DD-WRT as a security solution for a TP-Link AC1750 router, model Archer C7. In this article, we will explore deploying DD-WRT on a network as a basic solution, followed by an analysis of port scanning and vulnerabilities. We will also examine the ways in which DD-WRT help increase network security by installing additional functionality such as VPN, firewall, and other applications. We will study how an implemented DD-WRT solution will prevent cyber-attacks and vulnerability scans, also referring to research studies conducted to date. This paper provides a starting point for understanding the role of DD-WRT in enhancing IT security against cyber threats.

Speaker: Edi Timofte ("Ștefan cel Mare” University)

10:00 Malicious Short URLs Detection Technique

Nowadays, as the smartphone's performance grew a lot, being comparable with the one of computer systems, people are more attracted due to their capabilities and tend to use them in almost all activities. This fact determined the attackers to also consider the mobile devices among their targeted platforms. The easiest way for adversaries to get access to a smartphone is via SMS phishing (smishing), where they try to lure the victims into accessing a malicious URL to provide confidential data or download malware. To hide the features of these URLs and to avoid the text messaging's length restrictions, attackers use URL shortening services. In this paper, we propose an exhaustive system to detect malicious short URLs in smishing attacks by leveraging threat intelligence data from popular platforms like VirusTotal, and PhishTank, and by employing various Machine Learning (ML) algorithms that classify the SMS based on the features of the final redirected URL. Our system works for every URL, no matter the shortening service used either public or custom. Moreover, concerning the ML classifier, we took a publicly available balanced dataset for training, improved its feature set, and obtained an accuracy of approximately 97\%. The dataset contains 90 features that belong to three categories: the URL's lexical properties, external specifications, and the website content. We have tested our proposed ML model against JRip, PART, J48, and Random Forest algorithms, the last one being the most accurate. To showcase the effectiveness of our solution, we have implemented an Android app from scratch that detects malicious short URLs in SMS messages and notify the user concerning their legitimacy.

Speaker: Mr Răzvan Stoleriu (University Politehnica of Bucharest)

10:20 Coffee Break

10:40 Assisting students to compose music with deep neural networks and aesthetic measurement feedback

This paper introduces a novel approach that enables students to compose music using deep neural networks while simultaneously gaining insights into the functionality of various mathematically computed aesthetic measures. While there are existing applications that allow users to compose music with artificial intelligence, they often lack the feedback, which may be provided by newly developed mathematical aesthetic measurement techniques. Our approach bridges this gap by providing users with objective feedback for their generated melodies, fostering a better understanding of the underlying measurement methods. The application was evaluated by multiple participants, which made us aware of both the strengths and limitations of the application.

Speaker: Paroiu Razvan (UPB)

11:00 Approaches to teaching embedded development

TBA

Speakers: Mihai Chiroiu (University POLITEHNICA of Bucharest), Radu Mantu (University POLITEHNICA of Bucharest)

11:20 Improved Speech Activity Detection Model Using Convolutional Neural Networks

n this study, the performance of four speech/voice
activity detection (VAD) models, namely SpeechBrain, Picovoice,
InaSpeechSegmenter and WebRTC, is compared using data
collected from KAIST (Korea Advanced Institute of Science
and Technology). The goal is to develop an improved VAD
model based on the analysis of SpeechBrain’s performance and
provide a detailed perspective on the differences and similarities
between these models. The study focuses on the technologies
and algorithms used by each model and utilizes experimental
methods to evaluate their performance. The experimental results
show that SpeechBrain performs the best, with an average
recall value of 0.97 and a precision value of 0.96. Our research
endeavors have led to the refinement of the existing VAD model,
resulting in even more compelling performance metrics. The
improved model achieves a recall value of 0.98 and precision
value of 0.97, signifying its enhanced capability to accurately
detect speech activity. These findings hold promise for the future
advancement of VAD models and their application in various
speech-processing domains. This research can be used to enhance
future VAD models and develop more advanced speech processing
applications.
Index Terms—Improved Speech Activity Detection Model,
Convolutional Neural Networks (CNN), VAD (Voice Activity
Detection).

Speaker: Florin Pop (University Politehnica of Bucharest, Romania / National Institute for Research & Development in Informatics – ICI Bucharest, Romania)

11:40 Design and Implementation of a Distributed Firewall Management System for Improved Security

Network security is the practice of protecting computer networks and devices from unauthorized access, attacks, and threats. These threats can come from a variety of sources, including hackers, malware, and external attacks, and they can compromise the confidentiality, integrity, and availability of sensitive data and systems. A network security management system is a system that is designed to help protect and secure a network from these threats. In this context, this article presents an approach to create a management system for a distributed firewall, which provides multiple benefits that will be presented in the following sections. The script obtains the CPU utilization and network bandwidth data for each virtual machine and optimization algorithm, and then outputs the retrieved information. The application programming interface (API) is used to access the data from the virtualization platform and to modify the optimization process on the firewall. In this way, the solution is capable to monitor the entire network and make decisions based on parameters to ensure a high-performance security. The proposed script manages virtual machines inside the virtualization platform using an API to retrieve data and change the optimization algorithm of the firewall.

Speaker: Mr Andrei-Daniel Tudosi (USV)

Design and Implementation of a Distributed Firewall Management System for Improved Security.pptx

12:00 Low-code solutions integration for digital process automation - application in management

The research team of this article was involved in a publicly funded project implementation and has resorted to a low-code solution for digital process automation (Microsoft Power Platform), to boost project team efficiency. This research paper reveals the development process of the solution starting from the modeling phase to testing and implementation phase. The components included in the final solution are as follows: Microsoft Forms and SharePoint, Power Apps, Power Automate, and Power BI. All these combined resulted in a powerful solution for project implementation, monitoring, and reporting. The most important results achieved by implementing digital solutions were the following: reducing paperwork by eliminating repetitive tasks, increasing the motivation of the project team; improving personal digital skills; more efficient time management and more transparent and well-documented financial management.

Speaker: Mihai Timus (George Emil Palade University of Medicine, Pharmacy, Science and Technology of Targu Mures)

09:00 → 12:20 Session D

Session D

Convener: Dr Claudiu-Ionuț POPÎRLAN (University of Craiova)

09:00 VXLAN - A practical approach to cloud computing scalability

In this paper, we present a simple practical approach to using network virtualization technology in an ecosystem that is mainly composed of open-source software and operating systems.

Speaker: Claudiu Traistaru (Universitatea din Craiova)

09:20 Investigating User Characteristics and Behaviors on Social Media as Prequel for Fake News Detection

The rapid proliferation of social media platforms
has led to a remarkable increase of fake news dissemination,
creating a significant challenge for the integrity of online information.
To analyze this issue, this study aims to investigate the
relationship between user characteristics and behaviors on social
media platforms as a prequel for the development of effective fake
news detection strategies. For that purpose, we examined open
access datasets available for Romanian internet users and selected
several internet usage activities which we analyzed in relation
to key demographic characteristics that lead to establishing user
profiles prone to manipulations. By understanding the underlying
factors that contribute to the spread of fake news, we can design
targeted interventions to mitigate its impact and promote a more
informed and responsible online environment.

Speakers: Claudiu Ionut Popirlan (University of Craiova), Dr Cristina-Mihaela Tudorache (University of Craiova)

09:40 Mobile Application for Real-Time monitoring of an Air Quality Station

Abstract — The quality of the air around us is a major concern nowadays, a lot of health issues are associated with air pollution. This article presents an application of mobile and cloud technologies for real-time air quality monitoring. An Android mobile application was developed for monitoring the values of pollutants in the air from an air quality station. The station was designed to continuously acquire values from different sensors and to send them to databases. The mobile application provides real-time updates, notifications and historical values for the air parameters monitored.

Speaker: Gabriel-Catalin Nedianu (University of Craiova)

10:00 Hybrid Quantum-Classical Networks characteristics and optimization for Error Correction and Noise Mitigation

Hybrid Quantum-Classical Networks (HQCNs) have
emerged as a promising paradigm for harnessing the power of
both classical and quantum computing. This paper explores the
characteristics and optimizations associated with these networks.
Firstly, the characteristics of HQCNs are discussed, including the
integration of classical and quantum components, task allocation,
parallel processing, and the exploitation of quantum speedup.
The paper then delves into optimization techniques in hybrid
networks, such as dynamic task allocation, workload distribution,
and resource management. Optimizations for error correction
and noise mitigation in HQCNs involve fault-tolerant protocols
that enhance the resilience of quantum computations to errors
and noise. The paper concludes by emphasizing the importance of
understanding the characteristics and optimizations in HQCNs to
fully leverage their potential for solving complex computational
problems efficiently and effectively.

Speaker: Dr Claudiu Ionut Popirlan (University of Craiova)

10:20 Coffee Break

10:40 Automated script-based engine for Apache Kafka messaging system

Apache Kafka is a distributed event streaming platform used by thousands of companies worldwide for online data exchange. The system accepts any type of data, and this means that it can be used for varied types of applications. This paper presents an application that allows the user to send or receive different messages via Apache Kafka, either immediately or based on a schedule, using a Javascript-like language for programming.

Speaker: Lucian-Florentin Barbulescu (University of Craioba)

11:00 Sliding Window Method for Imputing Missing Data

The rapid technological advancement has led to an
increased amount of content and information which needs to
be reliably analyzed and correlated in order to make accurate
inferences. Data quality is a stringent issue when considering
gaps or missing information and various imputation methods,
algorithms or statistical techniques have been created to correct
missing data. To this purpose we propose an imputation method
using the Sliding Window Technique to help infer missing data
from datasets, a method that is efficient in terms of accuracy
and complexity. In order to illustrate the method, we present
a case study that simulates a scenario in which a half-yearly
dataset must be analyzed in relation to annual datasets. The
obtained results are passed to a statistical descriptive analysis
and the conclusions derived from the performed analysis suggest
that the estimations are homogeneous enough.

Speaker: Gabriel Stoian (University of Craiova)

11:20 Graphical Programming Environment for Creating Simulation Scenarios

Graphical Programming Environment is a software tool that allows developers to create, design, and manipulate code using visual elements, such as blocks, diagrams, and flowcharts. The proposed solution allows the creation for the simulation scenarios using visual blocks, and the execution of these scenarios on specific microcontrollers. This paper presents the stages of the environment: the configuration of the simulation tests, the generation of the source code, the compilation of the code, and the running of the generated executable to evaluate the results.

Speaker: Dr Eugen Ganea

11:40 The Significance of Enhancing Data Fairness in Educational Recommender Systems

Students, teachers, and educational institutions can all benefit from improving data fairness in educational RS through a variety of helpful approaches. Because data is so vital, the goal of this work is to propose characteristics and strategies to increase fairness in pre-processing procedures. By highlighting the importance of specific variables that are influenced by data fairness approaches, our work aims to improve data fairness in educational Recommender Systems (RS). We highlight the impact that enhancing data fairness has on students' performance in educational RS. Students' motivation rises and their performance benefits when they obtain fair and balanced recommendations. Additionally, we aim to present novel strategies that support increasing and strengthening fairness in educational RS.

Speaker: Alexandra Vultureanu-Albiși (University of Craiova)

12:00 Task Scheduling Policies in a Cloud Computing Environment

Virtualization solutions used within Cloud
computing have proven over time to address the problem of
inefficient use of physical computing resources. Virtualization
can provide high-level availability to critical applications with,
thus streamlining the operation of IT infrastructure and
responding quickly to changes. The paper proposes a
mathematical model based on Queuing Theory aimed at
ensuring a certain level of Quality of Service (QoS).
Keywords—cloud computing, queuing theory, task scheduling

Speaker: Catalina Mancas (Universitatea din Craiova)

12:30 → 12:50 Lunch