Research & Thesis topics – Francesco Musumeci

My research activity is focused on the broad area of telecommunication networks, with an emphasis on optical networking. Ongoing work is concentrated on the following areas:

Converged space-ground network infrastructure
Machine-Learning(ML)-assisted networking
Data-plane programmability
Optical networks architectures
Network resilience and survivability

Below, you can find more details for the ongoing activities on the various topics, with some suggested readings. MSc theses are currently available for each of these topics.

Do not hesitate to contact me for further details!

Current topics:

-) Converged space-ground network infrastructure

Future communications will provide high-speed connectivity to support extremely-low-latency and high-bandwidth services to literally any place in the world in a cost effective and ultra reliable manner. Existing networking technologies, i.e., optical, mobile-radio, satellite, etc., address these challenges while privileging some aspects over the others. For example, optical networks allow huge bandwidth utilization, but are difficult to be deployed in extreme ground geotypes (e.g., rural areas, mountain chains, oceans, poles). Conversely, satellite networks (e.g., Low Earth Orbit (LEO) constellations) enable extremely-wide Earth coverage but suffer from low available bandwidth. Realizing a Converged Space-Ground (CSG) infrastructure integrating terrestrial and satellite communication is a unique opportunity to reach reliable Internet connectivity anywhere and anytime without requiring additional abrupt investments.

Several Master Thesis are available in this context, concerning 1) the optimization of the converged space-ground network infrastructure, 2) the definition of novel protocol architectures allowing seamless data transfer between the two infrastructures.

-) Machine-Learning(ML)-assisted networking

Today’s telecommunication networks are sources of enormous amounts of widely heterogeneous data. This information can be retrieved from network traffic traces, network alarms, signal quality indicators, users’ behavioral data, etc. Then, mathematical tools based on Machine Learning are seen as a promising candidate to efficiently address optimization, management and design of telecommunication networks with such a great availability of data. Among these applications, it is worth mentioning the followings, but many others can be devised: failure management, traffic prediction, cybersecurity and privacy preserving, traffic classification, Quality of Transmission (QoT) estimation, etc.

-) Data-plane programmability & ML

Machine Learning (ML) models have found numerous applications in the automation of complex network management tasks. More recently, thanks to the introduction of new solutions for data-plane programmability (as the P4 pro- gramming language), it has become possible for programmable switches to execute ML-models directly in the data plane, with the great advantage that decisions can be taken at packet-rate, without the involvement of the control plane.

Research in this field (and available Master theses) concentrates on ML-assisted programmable data planes applied to various use cases, e.g., 1) cybersecurity, 2) application traffic classification, 3) resource allocation.

Suggested readings: JONS21, SIGCOMM21, ICC20

-) Optical networks design and management

Optical networks constitute the basic physical infrastructure of all large-provider networks worldwide, thanks to their high capacity, low cost and flexibility. Due to the high increase in the amounts of transported data, network operators are urged to devise efficient technical solutions to increase network capacity and improve its performance, while keeping costs under control, in terms of capital expenditures (Capex) and operational expenditure (Opex).

Here my work focuses on optimization of multi-band (C+L+S) optical networks, low-margin network design, filterless optical networks, optimized optical amplifier placement, and energy-efficient network architectures.

-) Network resilience and survivability

Disaster-based disruptions seriously degrading the performance of any communication network (e.g., due to natural disasters, technology-related disasters, or even malicious attacks) are now gaining importance due to observed increase of their intensity and scale. In presence of disasters, the unavailability of communication networks services, considered as an important part of the critical infrastructure in our society, implies evident societal problems, e.g., for people desperately seeking for information, or trying to communicate with each other. For these reasons, providing high resilience and availability to current communication networks is a key reasearch aspect nowadays.

Previous topic(s):

-) 5G, Centralized Radio Access Networks (C-RAN), and Network Function Virtualization (NFV)

5G networking represents a revolutionary concept of mobile access as it targets unprecedented performance, not only in terms of higher data rates per user, but also in terms of extremely-low latency, network automation and service capillarity.

To achieve this, 5G networks will resort to new technical solutions, including, e.g., small cell deployment, multipoint coordination (CoMP), Centralized Radio Access Network (C-RAN), content caching (e.g., for Video-on-Demand – VoD – delivery), Network Function Virtualization (NFV) and Software-Defined-Networks (SDN).