Day 1 03/12/2019
Room #1

Registration 08:30 - 09:00

Welcome and Opening Address 09:00 - 09:30

Dr. Max AGUEH - General chair, Welcoming Address by IMSP Authorities, Dr. Pélagie HOUNGUE – Local Chair

Keynote speech 1 09:30 - 10:30

Keynote - Pierre S. Dandjinou, ICANN, The Internet Domain Name System: Technology, Economics and Governance

Coffee break 10:30 - 11:00

Session 1 11:00 - 12:00

Security – Chair: Dr. Max AGUEH
11:00 - 11:00
Analysis of the impact of permissions on the vulnerability of mobile applications

In this paper, we explored the potential risks of authorizations unexplained by benign apps in order to maintain the confidentiality and availability of personal data. More precisely, we focused on the mechanisms for managing risk permissions under Android to limit the impact of these permissions on vulnerability vectors. We analyzed a sample of forty (40) apps developed in Burkina Faso and identified abuses of dangerous authorizations in several apps in relation to their functional needs. We also discovered combinations of dangerous permissions because it exposes the confidentiality of the data. This analysis allowed us to establish a link between permissions and vulnerabilities, as a source of risk of data security. These risks facilitate exploits of privileges that should be reduced. We have therefore proposed the need to coordinate resolution mechanisms to the administrators, developers, users to better guide the required permissions by benign apps on Android.
Authors: Gouayon KOALA (Université Joseph Ki-Zerbo (Burkina Faso)), Didier BASSOLE (Université Joseph Ki-Zerbo (Burkina Faso)), Aminata SABANE (Université Joseph Ki-Zerbo (Burkina Faso)), F. Tegawendé BISSYANDE (Université Joseph Ki-Zerbo (Burkina Faso)), Oumarou SIE (Université Joseph Ki-Zerbo (Burkina Faso)),
Hide Authors & Abstract

Show Authors & Abstract
11:00 - 11:00
On the Relevance of Using Multi-layered Security in the Opportunistic Internet-of-Things

Wireless Sensor Networks (WSNs) have recently gained more importance as key building blocks for the Internet of Things (IoT); a network infrastructure which has greatly increased in number of connected objects with instantaneous communication, data processing and pervasive access to the objects that we manipulate daily. However, WSNs may sometimes need to be deployed in an opportunistic fashion when there is no network with stable power supply available to support the dissemination of the sensor readings from their collection points to a gateway. For such deployments, traditional networking paradigms may fall short to secure the WSNs since most of the well-known security algorithms have been designed for the traditional high quality of service and fully connected networks. Building around some of the security algorithms and protocols which have been developed in the context of Delay Tolerant Networking, this paper presents a multi-layered security model for the opportunistic IoT. The model combines a hash based message authentication code (HMAC) algorithm implemented at the application layer of the IEEE 802.15.4 stack and an Access Control List (ACL) based identity based encryption algorithm used by the IEEE 802.15.4 MAC layer as a new and novel method of signing and authenticating data which is stored and forwarded on an opportunistic Internet of Things (IoT) infrastructure.
Authors: OLASUPO AJAYI (University of the Western Cape), Antoine Bagula (University of the Western Cape), Claude kakoko (University of the Western Cape), lutando Ngqakaza (University of the Western Cape),
Hide Authors & Abstract

Show Authors & Abstract
11:00 - 11:00
Analysis of Software Vulnerabilities Using Machine Learning Techniques

With the increasing development of software technologies, we see that software vulnerabilities are a very critical issue of IT security. Because of their serious impacts, many different approaches have been proposed in recent decades to mitigate the damage caused by software vulnerabilities. Ma-chine learning is also part of an approach to solve this problem. The main objective of this document is to provide three learning models predict effectively supervised for software vulnerabilities from a dataset of 6670 obser-vations. The effectiveness of the proposed models will be evaluated with several perfor-mance indicators including Accuracy.
Authors: Diako jerome (EDP INPHB Yamoussoukro, Côte d'Ivoire), ACHIEPO Yapo M. (Peleforo Gon Coulibaly University, Côte d'Ivoire), MENSAH Patrice (INPHB Yamoussoukro, Côte d'Ivoire),
Hide Authors & Abstract

Show Authors & Abstract

Lunch 12:00 - 13:30

Session 2 13:30 - 15:00

Smart Cities – Chair: Dr. Arnaud AHOUANDJINOU
13:30 - 13:30
A LoRaWAN coverage testBed and a multi-optional communication architecture for smart city feasibility in developing countries

Despite the inaccessibility or intermittency of Internet in developing countries, it remains one of the key elements in the field of IoT and the smart city. Most cities experiencing these difficulties can, therefore, use stand-alone LoRaWAN base station solutions and provide good coverage of the area. Leftovers can claim to have access to acceptable connectivity, and therefore cheaper wireless communication, giving direct access to Internet from the collection node, will be required for data transmission: Wi-Fi can attain these objectives. In this article, we first examine the existing tools to ensure a good coverage study of the LoRaWAN network and, thus, perform a testBed. After, we offer a multi-optional architectural model that supports LoRaWAN and Wi-Fi protocols and that is flexible for other options. A gateway (Wi-IoT) capable of both providing Wi-Fi access and at the same time the ability to collect, process and monitor data as a mini-server will be proposed as proof of concept. From the node to the gateway, the data is compressed and sent securely and a user who connects to the gateway can, then, have access to the data.
Authors: Pape Abdoulaye BARRO (Institute of Mathematics and Physical Sciences, African Center of Excellence in Applied Mathematics), Marco Zennaro (ICTP), Jules DEGILA (Institute of Mathematics and Physical Sciences(IMSP), Africa Center of Excellence in Mathematical Sciences and Applications, Porto-Novo, Benin),
Hide Authors & Abstract

Show Authors & Abstract
13:30 - 13:30
Effective Management of Delays at Road Intersections using Smart Traffic Light System

Rapid industrialization coupled with increased human population in urban regions has led to a rise in vehicle usage. The demand for space (road) by motorists for transportation has risen. Unfortunately, infrastructural development has not been at par with vehicular growth thus resulting in congestion along major roads. Traffic lights have been used for years to manage traffic flow. While they serve a good purpose, their underlining principle of operation is to a significant degree inefficient as traffic congestion still prevails and remains a major concern till date. This study seeks to tackle this challenge by proffering a Smart Traffic Management System (STMS) based on image detection. The system incorporates cameras which dynamically capture road situation as images, run them through an image processing algorithm to obtain traffic density then automatically adjust the service times at intersections. To measure the effectiveness of the approach, mathematical models were formulated, analytical comparison as well as experimental simulations were done. Results show that SMTS out-performed the Round-Robin algorithm used by traditional traffic lights, by reducing service interruptions, cutting delay times by at least 50\%, while remaining equally fair to all roads at the intersection. This system and its constituent components fall under the Edge computing paradigm as real time data capture, analysis and decisions are made by an embedded computer.
Authors: OLASUPO AJAYI (University of the Western Cape), ANTOINE BAGULA (University of the Western Cape), OMOWUNMI ISAFIADE (University of the Western Cape), AYODELE NOUTOUGLO (University of Lagos),
Hide Authors & Abstract

Show Authors & Abstract
13:30 - 13:30
Patterns for African smart cities semantic data integration

Smart cities is a topic broadly presented in the litterature. Numerous works have been done presenting the concept of smart city. But there is still a need concerning the matter of data integration in smart cities. In this paper we have presented a state of art of smart cities and semantic data integration.
Authors: Ferdinand Guinko (Department of computer science Institut Burkinabè des Arts et Metiers Université Joseph KI-ZERBO, Ouagadougou, Burkina Faso), Yaya Traoré (Université Joseph KI-ZERBO),
Hide Authors & Abstract

Show Authors & Abstract
13:30 - 13:30
Binary Search Based PSO for Master Node Enumeration and Placement in a Smart Water Metering Network

A Binary Search based Particle Swarm Optimization (BS-PSO) Algorithm is proposed for the enumeration and placement of Master Nodes (MNs) in a Smart Water Metering Network (SWMN). The merit of this proposal is it can simultaneously optimize the number of MNs as well as their locations. The Binary Search (BS) Mechanism searches a pre-specified range of integers for the optimal number of MNs. This algorithm iteratively invokes the PSO algorithm which generates particles based on the chosen number of MNs. The PSO uses these particles to determine MN coordinates in the fitness function evaluation process within the underlying SWMN simulation. The packet delivery ratio (PDR) is designated as the fitness value for the particle. Results for 10 BS-PSO optimization runs show that the median optimal number of MNs is 15 and that the mean PDR of 96 % can be realized. As part of future work, more optimization runs will be conducted to enhance the generalization of the results. The extension of this concept to other optimization algorithms such as Differential Evolution will also be considered.
Authors: Clement Nyirenda (University of the Western Cape), Samson Nyirongo (University of Namibia),
Hide Authors & Abstract

Show Authors & Abstract

Coffee break / Poster 15:00 - 15:30

Poster : Heterogeneous Sensor Network System for patients monitoring in intensive care units

Session 3 15:30 - 17:00

Internet & Internet of Things – Chair: Dr. Joel HOUNSOU
15:30 - 15:30
State of Internet Measurement in Africa - A Survey

This paper presents the results of a survey aimed at understanding the status of Internet measurement platforms usage, deployment and capabilities in Africa. It presents findings related to prevalence of measurement in the region, the reasons why the different business categories investigated conduct Internet measurement as well as the metrics of interest to these entities. The survey also looked at the popular measurement platforms that the respondents use in their measurement activities as well as the platforms that are hosted by businesses and users in the African region. The survey also recorded responses related to data handling and privacy considerations. A total of 123 responses were received from 34 countries. The survey revealed that Internet measurements are not widely conducted in the region largely due to the inadequacy of deployed measurement platforms, the lack of awareness in the subject, and the lack of relevant skills to carry out the measurement tasks. We outlined some recommendations to remedy these issues.
Authors: Musab Isah (AFRINIC), Amreesh Phokeer (AFRINIC), Josiah Chavula (University of Cape Town), Alemnew Sheferaw Asrese (Aalto University), Ahmed Elmokashfi (Simula Research Lab),
Hide Authors & Abstract

Show Authors & Abstract
15:30 - 15:30
I2PA, U-prove, and Idemix: An Evaluation of Memory Usage and Computing Time Efficiency in an IoT Context

The Internet of Things (IoT), in spite of its innumerable advantages, brings many challenges namely issues about users’ privacy preservation and constraints about lightweight cryptography. Lightweight cryptography is of capital importance since IoT devices are qualified to be resource-constrained. To address these challenges, several Attribute-Based Credentials (ABC) schemes have been designed including I2PA, U-prove, and Idemix. Even though these schemes have very strong cryptographic bases, their performance in resource-constrained de- vices is a question that deserves special attention. This paper aims to conduct a performance evaluation of these schemes on issuance and verification protocols regarding memory usage and computing time. Recorded results show that both I2PA and U-prove present very interesting re- sults regarding memory usage and computing time while Idemix presents very low performance with regard to computing time.
Authors: IBOU SENE (Ecole Polytechnique de Thiès), Abdoul Aziz CISS (Ecole Polytechnique de Thiès), Oumar NIANG (Ecole Polytechnique de Thiès),
Hide Authors & Abstract

Show Authors & Abstract
15:30 - 15:30
A Hybrid Network Model embracing NB-IoT and D2D Communications : Stochastic Geometry Analysis

A narrow-band system introduced in Release 13 of 3GPP has recently gain momentum to support a range of IoT use-cases. Narrowband-Internet of Things (NB-IoT) comes with low-cost devices characterized by extremely low power consumption, offering a battery life of more than 10 years, and broad radio coverage (up to 20 dB improvement over the standard LTE/GPRS systems) to target tens of kilometers, but on the cost of low data rate and higher end-to-end latency. NB-IoT can be deployed in three different operational modes; stand-alone as a dedicated carrier, in-band within the spectrum of a wideband LTE carrier, and within the guard-band of existing LTE carrier. In this paper, a tractable hybrid network model embracing both NB-IoT and D2D technologies has been introduced. we first present an analytical framework to derive analytical rate expressions for D2D in NB-IoT networks. Then, the performance gains of network model are investigated through numerical evaluations that demonstrate the superiority of proposed model over the traditional NB-IoT network.
Authors: Athanase ATCHOME (Laboratoire d’Electrotechnique, de Télécommunications et d’Informatique Appliquée ( LETIA), Abomey-Calavi, Bénin), Hafiz Husnain Raza Sherazi (Telematics Lab, Polytechnic University of Bari, Bari, Italy), Rodrigue Alahassa (Institut de Math\'ematiques et de Sciences Physiques(IMSP)/UAC, B\'enin), Frantz Tossa (ED-SDI), Dr. Thierry Edoh (Technical University of Munich), Luigi Alfredo Grieco (Department of Electrical and Information Engineering, Politecnico di Bari, Bari, Italy), Antoine Vianou (Ecole Doctorale Des Sciences de l'Ingénieur (ED-SDI)/UAC, B\'enin),
Hide Authors & Abstract

Show Authors & Abstract

Gala Dinner and Best Paper Announcement 18:30 - 22:00

Day 2 04/12/2019
Room #1

Keynote Speech 2, SEMECITY 09:30 - 10:30

SEMECITY: An African Ecosystem of Innovation

Coffee Break 10:30 - 11:00

Session 4 11:00 - 12:30

Data Management & IT Applications – Chair: Dr. Joel HOUNSOU
11:00 - 11:00
CLASSIFICATION OF PLANT SPECIES BY SIMILARITY USING AUTOMATIC LEARNING

The classification methods are diverse and vaiedy from one field of study to anoth-er. Among botanists, plants classification is done manually. This task is difficult, and results are not satisfactory. However, artificial intelligence, which is a new field of computer science, advocates automatic classification methods. It uses well-trained algorithms facilitating the classification activity for very efficient results. However, depending on the classification criterion, some algorithms are more efficient than others. Through our article, we classify plants according to their type: trees, shrubs and herbaceous plants by comparing two types of learning meaning the supervised and unsupervised learning. For each type of learning, we use the corresponding algo-rithms which are K-Means algorithms and decision trees. Thus we developed two classification models with each of these algorithms. The performance indicators of these models revealed different figures. We tested the models with a base of harvest-ed plants, and the results obtained are in accordance with the training results. Howev-er, for the K-Means, the total accuracy of the model is 81.16% with a margin of error of 18.83% and for the decision trees, we have an accuracy of 99.24% with a margin of error of 0.76%. We conclude that the decision trees are more appropriate in this type of classification.
Authors: ZACRADA FRANCOISE ODILE TREY (Institut National Polytechnique Houphouët-Boigny, Yamoussoukro,Cote d’Ivoire), GOORE BI TRA (Institut National Polytechnique Houphouët-Boigny, Yamoussoukro,Cote d’Ivoire), BROU MARCELLIN KONAN (Institut National Polytechnique Houphouët-Boigny, Yamoussoukro,Cote d’Ivoire),
Hide Authors & Abstract

Show Authors & Abstract
11:00 - 11:00
Mobile health applications future trends and challenges

Digital technologies are now a big part of the healthcare industry. With the rapid evolution of information and telecommunication technologies, mobile phones offers amazing opportunities to improve healthcare system, in the way that physicians, patients and other health system actors are more interconnected than it is in the traditional healthcare system. The goal of this study is to present the current and future trends in the field of mobile health (mhealth) and also to present the challenges in the use of mhealth apps.
Authors: Ivan Bessin (Department of computer science Institut Burkinabè des Arts et Metiers Université Joseph KI-ZERBO, Ouagadougou, Burkina Faso), Wilfried Ouedraogo (Department of computer science Institut Burkinabè des Arts et Metiers Université Joseph KI-ZERBO, Ouagadougou, Burkina Faso), Ferdinand GUINKO (Department of computer science Institut Burkinabè des Arts et Metiers Université Joseph KI-ZERBO, Ouagadougou, Burkina Faso),
Hide Authors & Abstract

Show Authors & Abstract
11:00 - 11:00
Laws and Regulations on Big Data Management: The Case of South Africa

A growing global trend has been witnessed in many developing countries where efforts and resources are been invested in advancement of electronic health information. The expectation is to improve the quality of health care, increase universal health coverage, and reduce both Legal Cases and healthcare costs in a changing world where data collected in the healthcare is producing big data sets which can provide useful insights for the advancement of healthcare. The challenge is a greater risk for legal regulations to keep up with the accelerated global changes resulting from Big Data, and loss of information privacy created by digital transformation. In some countries, legal, privacy and ethical issues related to use and access to personal health data still causes foreseeable challenges. This article will review the South African laws and regulations in handling, processing, storing, accessing and big data analytics on digital health data.
Authors: OLASUPO AJAYI (University of the Western Cape), Antoine Bagula (University of the Western Cape), Patrick Sello (University of the Western Cape),
Hide Authors & Abstract

Show Authors & Abstract
11:00 - 11:00
Big Data Processing Using Hadoop and Spark: The Case of Meteorology Data

Meteorology is a branch of science which can be leveraged to gain useful insight into many phenomenon that have a significant impact on our daily lives such as weather precipitation, cyclones, thunderstorms, climate change. It is a highly data-driven field that involves large datasets of images resulting from both radar and satellite, thus requiring ecient technologies for storing, processing and data mining to and hidden patterns in these datasets. Di erent big data tools and ecosystems, most of them integrating Hadoop and Spark, have been designed to address big data issues. However, despite its importance, only few works have been done on the application of these tools and ecosystems for solving meteorology issues. This paper proposes and evaluate the performance of a precipitation data processing system that builds upon the Cloudera ecosystem to analyse large datasets of images as a classification problem. The system can be used as a replacement to machine learning techniques when the classification problem consists of finding zones of high, moderate and low precipitations in radar images.
Authors: OLASUPO AJAYI (University of the Western Cape), Antoine Bagula (University of the Western Cape), Eslam Hussein (University of the Western Cape), Yahlieel Jafta (University of the Western Cape),
Hide Authors & Abstract

Show Authors & Abstract

Lunch 12:30 - 14:00

Session 5 14:00 - 15:00

AFRICATEK 2019 Workshop Session – Chair: Dr. Max Agueh
14:00 - 14:00
Africa’s Multilateral Legal Framework on Personal Data Security: What Prospects for the Digital Environment?

As the African continent continues to embrace technological innovations and corresponding infrastructures like the Internet of Things, concerns related personal data protection and security have been raised, which become more crucial as huge amounts of sensitive personal data are increasingly generated across the continent, especially with the proliferation of mobile banking. In response to these developments, African intergovernmental organizations have developed legal frameworks on personal data protection: ECOWAS has adopted a Data Protection Act, while the African Union (AU) has adopted a Convention on Cyber Security and Personal Data Protection. However, while other aspects of data protection law are more or less touched in these instruments, very little focus put on managing and safeguarding personal data security. This paper, in an attempt to present a critique of the state of affairs as regards personal data breach regulation in Africa, argues that the above African instruments do not provide a satisfactory response to current personal data privacy and security challenges Africa faces. Both instruments are significantly lacking in pre-breach and post-breach regulation, including breach reporting, liability rules and available remedies for affected data subjects. The paper concludes by recommending that these deficiencies be addressed in additional protocols to these instruments or in relevant future texts.
Authors: Rogers Alunge (Erasmus Mundus Doctoral Fellow, LAST-JD Program, CIRSFID, University of Bologna, Lungo Dora Siena 100A, Torino, Italy),
Hide Authors & Abstract

Show Authors & Abstract
14:00 - 14:00
AmonAI: a students academic performances prediction system

This paper presents a system, called AmonAI, that predicts the academic performances of students in the LMD system. The approach used allows to establish, for each of the teaching units of a given semester, some estimates of the students results. To achieve this, various machine learning techniques were used. In order to choose the best model for each teaching unit, we have tested 9 different algorithms offered by the Python Scikit-learn library to make predictions. The experiments were performed on data collected over two years at ”Institut de Formation et de Recherche en Informatique (IFRI)” of University of Abomey-Calavi, Benin. The results obtained on the test data reveal that, on five of the nine teaching units for which the work was conducted, we obtain an F2-score of at least 75% for the classification and an RMSE of less than or equal to 2.93 for the regression. The solution therefore provides relatively good results with regard to the dataset used.
Authors: Iffanice Houndayi (University of Abomey-Calavi (UAC)), Vinasetan Houndji (University of Abomey-Calavi (UAC)), Pierre Zohou (University of Abomey-Calavi (UAC)), Eugene Ezin (University of Abomey-Calavi (UAC)),
Hide Authors & Abstract

Show Authors & Abstract

Closing remarks 15:00 - 15:30