| Id | Project | Description | Link/Notes | Students |
|---|---|---|---|---|
| QoS Network Emulation with Realistic Traffic Types | Build an emulated network in Mininet, including video streaming, email/messaging, and file download servers to generate realistic application-level traffic. 3 servers (1 per type of traffic), 2 clients exchanging traffic with all servers and 3 switches (1 per client and 1 for all the servers). Configure a packet scheduling policy on the switches to prioritize streaming over download over emails. | Amir SANJAKDAR, Amir RAZZAQ ,Akash BALAMURUGAN, Jia Jie XU | ||
| Traffic Distribution Crafting and Evaluation of Scheduling Policies: DRR vs SP | Use Scapy to craft three different traffic distributions: Constant Bit Rate, Variable Bit Rate and Pareto distribution of packet size. Create a Mininet topology where three hosts send the generated traffic to a fourth host, through a switch. Compare the latency, packet loss and completion time of the traffic distributions using two scheduling policies in the switch: Deficit Round Robin and Strict Priority Scheduling (arbitrary parameters) | ZERRAD Adan, EL GARAH Itminene, SIVANESAN Stéphane, N'DIAYE Waly | ||
| Traffic Distribution Crafting and Evaluation of Scheduling Policies: DRR vs Shaping | Use Scapy to craft three different traffic distributions: Constant Bit Rate, Burst Traffic and Exponential Distribution of inter-arrival time. Create a Mininet topology where three hosts send the generated traffic to a fourth host, through a switch. Compare the latency, packet loss and completion time of the traffic distributions using two scheduling policies in the switch: Deficit Round Robin and Traffic Shaping (arbitrary parameters) | ASSOUL Lydia, BANTIKOS Nicolas, Aldea Palamani | ||
| Containerisation of real-time federated Learning agents with Kafka | Containerizing data pipeline system and the federated learning functions using docker containers | Data pipeline system and functions to containerize: https://gitlab.roc.cnam.fr/data-pipeline-system-designs-for-in-network-learning function description in the paper: https://hal.science/hal-04681121 |
AMEUR Samia, GUEDAIEM Marwan, MEDDAH Khaled Fouad, BENELMIR Rami | |
| Containerisation of real-time federated Learning agents with RabbitMQ | The current data pipeline system is based on the publish/subscribe model implemented with Apache Kafka. Replace the Apache Kafka by RabbitMQ and containerize the functions of the data pipeline system and the federated learning functions using docker container. | RabbitMQ: https://www.rabbitmq.com/ Data pipeline system and functions to containerize: https://gitlab.roc.cnam.fr/data-pipeline-system-designs-for-in-network-learning function description in the paper: https://hal.science/hal-04681121 |
KHAN Muneeb, GHOUT Sofiane, GABLI Wassif-Eddine, SOILIHI Djouhoudi | |
| Containerisation of real-time federated Learning agents with ZENOH | The current data pipeline system is based on the publish/subscribe model implemented with Apache Kafka. Replace the Apache Kafka by ZENOH and containerize the functions of the data pipeline system and the federated learning functions using docker container. | ZENOH: https://github.com/eclipse-zenoh/zenoh Data pipeline system and functions to containerize: https://gitlab.roc.cnam.fr/data-pipeline-system-designs-for-in-network-learning function description in the paper: https://hal.science/hal-04681121 |
Sait Tinhinane, Ismail Mustafa, KOLINGBA Aaron, Iratni Mounira | |
| Comparison of packet delivery ratio (PDR) for LoRaWAN with and without FHSS option (Direct to satellite) | This project compares the packet delivery ratio (PDR) of LoRaWAN with and without Frequency Hopping Spread Spectrum (FHSS) to evaluate its impact on communication reliability in IoT applications when connecting to the satellite. | https://github.com/MuhammadAsadUllah1/Analysis-and-Simulation-of-LoRaWAN-LR-FHSS |
ABDELHAK Razi, ALEXANDRE KIDNEY PERKS, MOHAMED Asri, EMEL Tatoglu | |
| Impact of IoT application data rate on the performance of the network | IEEE 802.15.4-TSCH is the foundational PHY and MAC layer for many time-critical industrial IoT technologies, such as ZigBee and WirelessHART. TSCH combines TDMA and FDMA in its MAC layer to ensure high reliability. In TSCH-Sim, change the data traffic of nodes in a given topology scenario from very light (one packet every minute) to very heavy (one packet every 500 milliseconds) to observe how network performance degrades. The simulation can be conducted for topologies with 10, 20, 40, 80, 160, and 320 nodes. | https://github.com/edi-riga/tsch-sim For those prefer C programing can use Contiki-ng instead of Tsch-sim: https://docs.contiki-ng.org/en/master/doc/tutorials/TSCH-and-6TiSCH.html |
Hibat allah ZINE EL AABIDINE, Hayoub MOUFFOK, Nuno SOARES, Amordei BOLENGE, Ismail YADEN | |
| Malware detection with supervised learning | Use CTU-13 dataset to train an ML/DL model to detect Malware (data preprocessing is required) | https://mcfp.felk.cvut.cz/publicDatasets/CTU-Malware-Capture-Botnet-42/detailed-bidirectional-flow-labels/capture20110810.binetflow | MAOUCHE Naim, ZOLFAGHARI MOGHADDAM Karaneh, IBRAIMO Nacir, PANCHENKO Kristina | |
| Deploy a Honeypot for Network Threat Analysis – ssh and dns | Deploy a real-world honeypot on a cloud server. Setup a system featuring a ssh and dns server. Log every incoming connection (e.g., IP address, geographical location, time of day, attack patterns, payloads, traffic volume, login attempts) in a parsable format (e.g., json). | @lecnam.net emails should grant a free demo account for students on Azure. | ICHALAL Smail, KHRISSI Hamza, MOUBAYED Ayla, SALEH Sami | |
| Deploy a Honeypot for Network Threat Analysis – http(s) and ftp | Deploy a real-world honeypot on a cloud server. Simulate a vulnerable system featuring an http(s) and ftp server. Log every incoming connection (e.g., IP address, geographical location, time of day, attack patterns, payloads, traffic volume, login attempts) in a parsable format (e.g., json). | @lecnam.net emails should grant a free demo account for students on Azure. | ARULANANTHAM Mekkhishan, ALAOUI BELGHITI Hanaa, MENDES VINTENA Victor, JEAN-BAPTISTE Dorian | |
| Container migration – Docker vs Podman | Conduct a comparative study of live container migration between two servers using CRIU with Docker and Podman. Examine the differences in downtime, resource usage, and the ease of setup. | https://criu.org/Main_Page | ABBASSEN mouhcene, DJILI ousama, AMZALI mouloud, Bianca Catarau | |
| Container migration – cri-o vs Docker | Conduct a comparative study of live container migration between two servers using CRIU with Docker and cri-o. Examine the differences in downtime, resource usage, and the ease of setup. | https://criu.org/Main_Page | FREJABUE Anaïs, YAHIATENE Ghiles, NGUYEN Ngoc Anh, KHORSI Rezak | |
| Traffic Engineering in SDN Using OpenFlow | Create a Mininet SDN environment with Open vSwitch and an ONOS controller. The topology should include at least 5 nodes and 10 hosts. Implement a dynamic flow rerouting traffic engineering policy based on link utilization. Evaluate the performance impact on latency, throughput, and link efficiency. | CLAIR Arnaud, BOUDJEMAHI Meriem, SITANG Ninine, LACHEMOT Cédric | ||
| Performance Evaluation of QUIC vs TCP | Set up a virtual environment (locally or in the cloud) consisting of a client machine and 2 servers: one configured for QUIC and the other for TCP connections. Simulate different network conditions between the client and the servers, in terms of pkt loss, latency and bit rate. Compare the performance of the two protocols in terms of throughput, latency and handshake completion time. | QUIC implementation for servers: https://github.com/cloudflare/quiche @lecnam.net emails should grant a free demo account for students on Azure. |
BESSABIS Abdeldjaoued, KABTANE Rania, MOUSSOUNI Yaakoub, BOUROUCHA Abdelatif. | |
| UE Mobility in 5G Networks with Open5GS and srsRAN | Simulate a 5G network using Open5GS as the core network and srsRAN as the RAN. Deploy one UE and 2 gNBs, and simulate the UE moving from the coverage area of one GNb to the other. Measure the handover completion time, as well as the throughput before, during and after the handover. | SrsRAN: https://github.com/srsran Open5GS: https://github.com/open5gs |
ROSSIGNOL Paul, SMIDA ZOUINA ROFAIDA, NGUYEN HUU TRUC, FAYE ABDOU AZIZ | |
| UE Mobility in 5G Networks with FREE5GC and OpenAirInterface | Simulate a 5G network using FREE5GC as the core network and OpenAirInterface as the RAN. Deploy one UE and 2 gNBs, and simulate the UE moving from the coverage area of one GNb to the other. Measure the handover completion time, as well as the throughput before, during and after the handover. | FREE5GC: https://github.com/free5gc/free5gc OpenAirInterface RAN: https://openairinterface.org/oai-5g-ran-project/ https://gitlab.eurecom.fr/oai/openairinterface5g/ |
Myriame BCHIR, Ena IDDIR, Katia TERRACHET, Antoine RAYMOND | |
| UE Mobility in 5G Networks with ns-3 LENA 5G Simulator | Simulate a 5G network in ns-3, using the Evolved Packet Core module as the core network and the LENA 5G module as the RAN. Use the mobility models available in ns-3 to move one UE in the simulation area, recreating two scenarios: linear path through gNB areas, and circular path around a cluster of 3 gNBs. Measure the handover completion time, as well as the throughput before, during and after the handover. | LENA 5G: https://cttc-lena.gitlab.io/nr/html/ |
Doroteja Petrovic, FOREST Salim, Dysan Nebugath-nacher | |
| Comparative Study of SDN Controllers: Faucet, ONOS, and OpenDaylight | Set up and evaluate three SDN controllers (Faucet, ONOS, ODL) using a Mininet network with multiple OpenFlow switches arranged in a 3-tier topology. Compare the controllers’ performance in terms of latency and goodput (intended as the ratio between the successfully transmitted user traffic to total traffic, including packet loss and control plane overhead). Conduct the evaluation under increasing traffic loads and number of switches. Implement additional features such as telemetry and QoS policies, assessing the ease of deployment for each controller. | Faucet: https://faucet.nz | BENDAKIR Aymen, ALLOUACHE Ayhem, ZORGANI Rayene Nassim | |
| Emulating Lifecycle Management of Virtualized Network Functions | Demonstrate the use of the VNF LCM Emulator, developed by the ETSI Industry Specification Group for Network Functions Virtualization (ISG NFV). | Article: https://superuser.openinfra.dev/articles/emulating-lifecycle-management-of-virtualised-network-functions/ Tool: http://tools.etsi.org/vnf-lcm-emulator/ |
BOUCHEFIRAT Loukmane, BOUTEBICHA Ihssan, FELIH Riad, FERRAHI Kamel-soaid | |
| BPP packet trimming Demo | Demonstrate the use of the packet trimming technique to implement In-Network Video Quality Adaption. | Papers: https://discovery.ucl.ac.uk/id/eprint/10171122/1/icin_2023.pdf https://dl.acm.org/doi/pdf/10.1145/3458305.3478440 |
MESBAHI Said, SEHAKI Thamila, SLAMANI Lila, DAOUI Mokrane Walid | |
| P4-UPF | Demonstrate the implementation of a 5G core UPF on a P4 switch, using the BMv2 software switch. | https://dl.acm.org/doi/pdf/10.1145/3482898.3483358 | HAMMOUDI Nasreddine, BA Sokhna Khadijatou | |
| NWDAF OAI Demonstration | Demonstrate the Network Data Analytics Function features using the OpenAir Interface implementation | https://gitlab.eurecom.fr/oai/cn5g/oai-cn5g-nwdaf https://gitlab.eurecom.fr/oai/cn5g/oai-cn5g-nwdaf/-/blob/master/docs/TUTORIAL.md |
Walid Berrouk, Lyna Ait Kaci, Imene Silakehal, Yasmine Si Chaib | |
| Omnipaxos for replication and fault-tolerance in Virtual Network Functions | Demonstrate the features of Omnipaxos to handle replication and fault-tolerance of a system featuring load balancing and firewall VNFs deployed over a scalable number of VMs (or containers). Showcase the response of the system to the events of network failure and abrupt VM shutdown. | https://omnipaxos.com/ https://github.com/haraldng/omnipaxos/tree/master |
ELMOKRETAR Nour elhouda nada, MOUINE Youssef, TAHIR Issam, Zhong Yufei | |
| Threat Assessment & Response System AI model for automated penetration testing | Demonstrate the use of TARS AI Agents system to perform cybersecurity penetration testing of at least 5 websites of your choice. Assess the capability of the system and comment on the major vulnerabilities found and potential countermeasures. | https://github.com/osgil-defense/TARS | ARCHA Aya, AMENTAK Nouhaila, DANE Mohamed, SEKHRI Abdelali | |
| Synthetic Data Generation Using Machine Learning - MAWI | Implement a generative ML model to create new synthetic data from the MAWI Archive dataset. Assess the quality of the generated data points by providing at least 5 metrics of fidelity and diversity (e.g., comparison of mean/std deviation, KNN). | MAWI Archive: https://mawi.wide.ad.jp/mawi/ Telecom GPT | SAMOURA Amadou, Hacene Katia, Jérémy NGUYEN | |
| Synthetic Data Generation Using Machine Learning - 5G3E | Implement a generative ML model to create new synthetic data from the 5G End-to-End Emulation dataset. Assess the quality of the generated data points by providing at least 5 metrics of fidelity and diversity (e.g., comparison of mean/std deviation, KNN). | 5G3E dataset: https://github.com/cedric-cnam/5G3E-dataset. Telecom GPT | Ahmad Jaber, Oladipupo Kola, Hassan Chreif, Ahmad ABDALLAH |