How much time is required to complete the thesis?
The minimum work to complete a master's thesis is six months full-time, and three months full-time to complete a tirocinio/internship (bachelor thesis). Full-time means 8 hours/day x 5 days/week. Hence if your peace is 50%, you will need 6 and 12 months, respectively.
When should a student ask for the assignment?
Ask the thesis when you know you can work on the project at least 50% of your time. It is reasonable to start working on the thesis/internship/tirocinio when only two exams are left.
Should a student provide some document/info when requesting the thesis?
You must send me your academic record so I can check your competencies. Typically, working on a subject without the required background is a vast of time for the student and the advisor, especially for internship/tirocinio.
Can I perform AFC (Complementary Formative Activities, 6 CFU, Computer Science program) on the proposed projects?
Yes, the proposed projects are also intended for AFC.
Are there prerequisites for thesis/internship/tirocinio on Cybersecurity?
Many students aim to do a thesis, internship, or tirocinio, on Cybersecurity, one of the most complex research/professional fields. Unfortunately, not all students can handle such complexity, as verified in recent years.
Prerequisites to complete a thesis, internship, or tirocinio, in Cybersecurity are the following:
good practical skills in C programming
good knowledge of Operating systems (from both a theoretical and practical perspective) and Networking
good background in math and algebra
a grade >= 28/30 for the Sicurezza exam (or any other exam on Cyber Security fundamentals)
This thesis work is intended for highly motivated students with strong academic skills and a solid study record. It is in collaboration with Storm Reply Roma - Students will work directly with the company (Suggestion - work suitable for a Master's Thesis)
This thesis presents a comprehensive investigation into the application of generative AI technologies, specifically Amazon Q, for accelerating and optimizing cloud migration strategies on Amazon Web Services. The research explores the theoretical foundations and practical implementations of AI-driven migration methodologies, examining how large language models and generative AI capabilities can transform traditional migration approaches through intelligent automation, pattern recognition, and predictive optimization.
The study analyzes Amazon Q's architecture and its integration with AWS migration services, evaluating its effectiveness in automating critical migration phases including discovery and assessment, migration planning, and post-migration optimization. Through quantitative metrics, the research measures the impact of AI-assisted migration on key performance indicators such as migration velocity, error reduction rates, downtime minimization, and overall project timeline compression.
The experimental component implements real-world migration scenarios across diverse workload types, including legacy monolithic applications, microservices architectures, and data-intensive systems. The research documents how Amazon Q's generative capabilities facilitate automated code modernization, infrastructure-as-code generation, and intelligent resource right-sizing. Special attention is given to the AI's ability to generate migration runbooks, identify anti-patterns, and provide contextual recommendations based on AWS best practices.
This thesis work is intended for highly motivated students with strong academic skills and a solid study record. It is in collaboration with Storm Reply Roma - Students will work directly with the company (Suggestion - work suitable for Bachelor's and Master's Theses)
This thesis presents a comprehensive analysis and implementation of three distinct container image signing methodologies for Amazon Elastic Kubernetes Service (EKS) deployments: traditional asymmetric key-based signing, AWS Signer managed service approach, and Sigstore Cosign keyless signing framework. The research investigates the security architectures, operational workflows, and trust models underlying each approach, evaluating their effectiveness in ensuring container image integrity and authenticity within cloud-native environments.
The study examines critical security aspects including signature verification mechanisms, key management lifecycles, certificate chain validation, and integration patterns with existing CI/CD pipelines. Through quantitative analysis, the research evaluates each approach's performance metrics including signing throughput, verification latency, scalability limits, and operational overhead in production-grade EKS clusters.
The experimental component implements end-to-end image signing workflows for all three approaches, incorporating automated policy enforcement through admission controllers and continuous compliance monitoring. Special attention is given to supply chain security implications, SLSA (Supply chain Levels for Software Artifacts) compliance levels, and integration with container registries and vulnerability scanning tools. The analysis includes total cost of ownership calculations, security posture assessments, and operational complexity evaluations.
This thesis work is intended for highly motivated students with strong academic skills and a solid study record. It is in collaboration with Storm Reply Roma - Students will work directly with the company (Suggestion - work suitable for Bachelor's and Master's Theses)
This thesis presents a systematic comparative analysis of three leading Kubernetes management platforms: Rancher, VMware Tanzu, and Amazon EKS Everywhere. The research investigates the architectural foundations, deployment models, and operational characteristics of each platform, evaluating their effectiveness in orchestrating containerized workloads across hybrid and multi-cloud environments.
The study examines critical aspects including cluster provisioning mechanisms, security frameworks, monitoring and observability capabilities, and integration patterns with existing enterprise infrastructure. Through quantitative metrics and qualitative assessments, the research evaluates each platform's performance in terms of scalability, resource utilization, deployment complexity, and operational overhead.
The experimental component implements representative use cases across all three platforms, conducting performance benchmarking and stress testing under various workload scenarios. Special attention is given to multi-cluster management capabilities, GitOps integration, and disaster recovery mechanisms. The analysis includes total cost of ownership (TCO) calculations and return on investment (ROI) projections for different organizational scales.
This thesis work is intended for highly motivated students with strong academic skills and a solid study record. It is in collaboration with Storm Reply Roma - Students will work directly with the company (Suggestion - work suitable for Bachelor's and Master's Theses)
This thesis compares three Kubernetes networking approaches on Amazon EKS: Cilium (eBPF data plane with L3–L7 policy and deep observability), AWS VPC CNI (native VPC integration using ENIs for low latency and predictable routing), and Multus (meta‑CNI enabling multiple Pod interfaces for specialized planes like SR‑IOV/DPDK). The study evaluates architecture, performance (throughput, P50/P99 latency, CPU/memory overhead), scalability (pod density, IP management, ENI limits), security (policy depth, identity awareness), and operations (install/upgrade, troubleshooting, cost). Methodology includes reproducible benchmarks (iperf3/Fortio), chaos/failure tests, and dual‑stack scenarios across small/large clusters.
With more data being generated from thousands of satellites, especially constellations, it makes sense to extend edge computing, the cloud, and ML/AI capabilities to orbit, especially if it proves to be more reliable and drives costs down. The European Commission is even looking into the possibility of moving entire data centers to orbit, reducing carbon emissions from data centers on Earth [1]. For example, moving ML/AI tasks on satellites for in-orbit analysis and processing will reduce data downlink [2].
A space Cloud comprises thousands of micro-datacenters installed in the constellation's satellites, forming a distributed cloud. Scheduling tasks for satellite operation is a challenge considering tasks should be executed in a specific time frame with stringent deadlines. Limited resources and energy budget make harder the problem.
This project aims to develop distributed task-scheduling algorithms for the space Cloud.
Research contracts for brilliant students are available
In this project we are investigating how to build self-protecting systems, that is computer systems and applications capable to autonomously detect and respond to cyber attacks.
The project is founded by the Italian Ministry of Research in the framework of the PRIN 2022 program.
In this project you can contribute to the development of a simulated cyber-range, that is a platform to simulates cyber attacks and technics for automatic detection and autonomous response to the attacks.
Research contracts for brilliant students are available
--- This project is discontinued ---
In this project, we are developing a next-generation cloud service broker helping the transition of the public administration to the cloud (PNRR objective)
The service broker supports the following functionalities
Automatic generation of technical specification from human-oriented specification
Service discovery
Service recommendation
QoS monitoring and Assurance
Cost monitoring and prediction
Students can contribute to the development of one of the above functionalities or their integration
--- We successfully completed this project. No more thesis available ---
Energy communities organise collective and citizen-driven energy actions that help pave the way for a clean energy transition while moving citizens to the fore. They contribute to increasing public acceptance of renewable energy projects and make it easier to attract private investments in the clean energy transition. At the same time, they have the potential to provide direct benefits to citizens by increasing energy efficiency, lowering their electricity bills and creating local job opportunities [https://energy.ec.europa.eu/topics/markets-and-consumers/energy-communities_en]
In this project, we will analyze social networks like Meta, X, and Telegram to discover the presence of Energy communities, and we will explore their contents and dynamics.
Specifically, students are requested to build a dataset by collecting data from the internet and to perform the analysis using NLP and data mining tools and methodologies.
Research contracts for brilliant students are available