PhD Position F/M PhD : Energy efficient slicing for 6G networks using AI/ML technics

The Inria centre at Université Côte d'Azur includes 37 research teams and 8 support services. The centre's staff (about 500 people) is made up of scientists of different nationalities, engineers, technicians and administrative staff. The teams are mainly located on the university campuses of Sophia Antipolis and Nice as well as Montpellier, in close collaboration with research and higher education laboratories and establishments (Université Côte d'Azur, CNRS, INRAE, INSERM ...), but also with the regiona economic players.

With a presence in the fields of computational neuroscience and biology, data science and modeling, software engineering and certification, as well as collaborative robotics, the Inria Centre at Université Côte d'Azur  is a major player in terms of scientific excellence through its results and collaborations at both European and international levels.

Within the framework of a partnership

• collaboration between 2 Inria teams: neo and coati and a company Nokia,

With the current energy crisis, the skyrocketing cost of energy, and the global awareness of the
consequences of carbon emissions on our planet, it is crucial to reduce the global energy
consumption related to Information and Communication Technologies, and in particular the one
related to networks.

Indeed, while being commercialized worldwide, the 5G technology coupled with ultra high
resolution video has been blamed for its high energy consumption [5G-power]. To answer this
issue, industrials and researchers have started to look beyond 5G to define the next 6G with at
its core the need to evolve towards greener networks. The 6G standard imposes a
transmission energy efficiency target in its Key Parameter Indicators (1picoJ/bit). However,
at the same time, 6G will introduce several technological breakthroughs with the integration of
traditional terrestrial mobile networks with emerging space, aerial, and underwater networks to
provide anytime, anywhere, network access. Another critical paradigm of 6G is the utilization of
Artificial Intelligence (AI) techniques to provide context-aware information transmissions and
personal-customized services, as well as to realize automatic network management [roadmap].
The growing ICT infrastructure, exploding data, and the AI-based services will result in surging
energy consumption.

Thus, succeeding in the challenge of developing more energy efficient networks will require
significant improvement in several directions [survey-green]: e.g. re-launching measurement
campaigns, rethinking protocols, developing energy-efficient network management algorithms
[algo], adopting energy harvesting techniques [harvesting], deciding if and when data should be
sent, but also creating tools to allow the individuals to take informed decisions and have a
sustainable Internet usage, in particular with the high definition video traffic.

We will focus on the provisioning and management of energy efficient network services.

With the advent of next generation networks implementing Software Defined Networks (SDN)
and Network Function Virtualization (NFV), network services can be set up dynamically at the
right time and at the right place. It has been shown that SDN can improve the energy efficiency
of networks by allowing fast rerouting. This allows to turn off redundant parts of the networks
used for backup in case of failures or over-dimensioning in case of sudden peaks of traffic
[sdn-energy] but also to reconfigure the network when the demand has changed
[reconfiguration]. The use of virtualization also reduces the use of specialized and
energy-intensive middleboxes and to provision network and data center resources only whenand where needed [nfv-energy].

However, with the multiplication of usages, connected cars, cities, factories, more generally with
the development of the Internet of Things, the number of services a network has to handle
increases drastically. At the same time, with the consideration of micro-services composed to
build network functions which are then added to form network services, the size of the latter also
increases. New methods have to be considered to be able to provision them on the fly in a fast
and scalable way, while not increasing network energy consumption.

During the PhD we will tackle theses problems following the guidelines written below:
- First, we plan to refine existing power models as there are the foundations to discuss
energy efficiency. We will make a special focus on video services, as video represents
the majority of the Internet traffic and drives a significant share of the investment in terms
of data centers and Internet links. We will start from the methods proposed in
[power-models] proposed methods to estimate the carbon footprint of a typical streaming
service.

- Second, we plan to explore how to use new AI methods for this challenge. AI/ML
methods are envisioned as a way to progress towards green networks [AI-for-green]. In
particular, new learning methods have been proposed to solve large scale combinatorial
problems, some of which are at the core of network placement problems. As an
example, [NIPS-billion] presents a reinforcement learning method to solve covering
problems for billions-sized graphs. This RL method has a good potential to help
provisioning network services, as placing services can be seen as a covering problem,
see [covering]. We will thus investigate how to use these new AI methods to provide
approximate solutions for provisioning a very large number of network services.

- Third, we plan to study how AI models can help to reconfigure the networks efficiently
when (i) demand or (ii) energy availability has changed. (i) First, the level and nature of
traffic strongly vary during the day. Second, with the development of smart cars, UAVs,
extra-terrestrial networks, there will be a high mobility of end-nodes. Network services
will thus have to be sufficiently instantiated, reconfigured, and stopped within a short
time scale. (ii) Indeed, energy harvesting has been widely recognized as an important
part for green communications. Part of such energy resources is uncontrollable but
predictable (such as solar, winding, tide, and other renewable sources). Another part is
partially controllable such as Radio Frequency energy harvesting [harvesting]. Another
issue that will be faced during next winters in Europe concerns planned power cuts. The
challenge is to find ways to manage the networks and the expected demands with such
varying sources of energy? We will study how AI models using energy predictions can
allow reconfiguring the network in advance to adapt to the changing energy availability,
e.g. to choose when doing a large backup, to plan the switching off of some network
parts.

References.
[roadmap] Letaief, Khaled B., et al. (2019) "The roadmap to 6G: AI empowered wireless
networks." IEEE communications magazine.
[harvesting] Chu, M., Li, H., Liao, X., & Cui, S. (2018). Reinforcement learning-based
multiaccess control and battery prediction with energy harvesting in IoT systems. IEEE Internet
of Things Journal.
[5G-power] “HUAWEI:5G Power white paper,”Shenzhen, China, HuaweiTechnol. Co., Ltd.,
White Paper. (2020). [Online]. Available:
https://carrier.huawei.com//media/CNBG/Downloads/Spotlight/5g/5G-Power-White-Paper-en.pdf
[algo] Vallero, G., Renga, D., Meo, M., & Marsan, M. A. (2019). Greener RAN operation through
machine learning. IEEE Transactions on Network and Service Management.
[survey-green] Huang, T., Yang, W., Wu, J., Ma, J., Zhang, X., & Zhang, D. (2019). A survey on
green 6G network: Architecture and technologies. IEEE access,.
[AI-for green] Mao, B., Tang, F., Kawamoto, Y., & Kato, N. (2021). AI models for green
communications towards 6G. IEEE Communications Surveys & Tutorials.
[energy-sdn] Huin, N., Rifai, M., Giroire, F., Pacheco, D. L., Urvoy-Keller, G., & Moulierac, J.
(2018). Bringing energy aware routing closer to reality with SDN hybrid networks. IEEE
Transactions on Green Communications and Networking, 2(4).
[virtualization-sdn] Huin, N., Tomassilli, A., Giroire, F., & Jaumard, B. (2018). Energy-efficient
service function chain provisioning. Journal of Optical Communications and Networking.
[reconfiguration] Gausseran, A., Giroire, F., Jaumard, B., & Moulierac, J. (2021). Be scalable
and rescue my slices during reconfiguration. The Computer Journal.
[NIPS-billion] Manchanda, S., Mittal, A., Dhawan, A., Medya, S., Ranu, S., & Singh, A. (2020).
Gcomb: Learning budget-constrained combinatorial algorithms over billion-sized graphs.
Advances in Neural Information Processing Systems (NIPS)..
[cover] Tomassilli, A., Giroire, F., Huin, N., & Pérennes, S. (2018). Provably efficient algorithms
for placement of service function chains with ordering constraints. In IEEE Conference on
Computer Communications (INFOCOM).
[power-models] Stephen Makonin, Laura U. Marks, Radek Przedpelski, Alejandro
Rodriguez-Silva, Ramy ElMallah.Calculating the Carbon Footprint of Streaming Media: Beyond
the Myth of Efficiency. LIMITS: Eighth Workshop on Computing within Limits 2022.
[harvesting] Kansal, A., Hsu, J., Zahedi, S., & Srivastava, M. B. (2007). Power management in
energy harvesting sensor networks. ACM Transactions on Embedded Computing Systems
(TECS).

Research

Networking

Algorithms

Optimization

AI/Machine learning

• Subsidized meals
• Partial reimbursement of public transport costs
• Leave: 7 weeks of annual leave + 10 extra days off due to RTT (statutory reduction in working hours) + possibility of exceptional leave (sick children, moving home, etc.)
• Possibility of teleworking (after 6 months of employment) and flexible organization of working hours
• Professional equipment available (videoconferencing, loan of computer equipment, etc.)
• Social, cultural and sports events and activities
• Access to vocational training
• Social security coverage

Duration: 36 months
Location: Sophia Antipolis, France
Gross Salary per month: 2082€ brut per month (year 1 & 2) and 2190€ brut per month (year 3)