Seminar on Topics in Communications Networking

Lecturer (assistant)
Number0000001067
TypeAdvanced seminar
Duration3 SWS
TermWintersemester 2020/21
Language of instructionEnglish
Position within curriculaSee TUMonline
DatesSee TUMonline

Dates

Admission information

See TUMonline
Note: Participation in the Kick-Off lecture (first lecture - see TUMonline) is mandatory to get a topic assigned. Prior registration in TUMonline is also mandatory. In case of more student registrations than topics available, topics assignment will follow a predefined selection process.

Objectives

The goal is to teach skills in literature review, scientific writing and presentation. The seminar is intended to introduce students to current research literature to bridge the gap between coursework and thesis research. The student will gain understanding of a particular research area, learn how to apply the understanding, analyze the results, and evaluate their impact.

Links

Topics (WS 20/21)

If you are interested to do your seminar on one of the topics offered by LNT or LKN, kindly contact the respective supervisor to show your interest (including the motivation to select this topic). If you reach an agreement with the respective supervisor before 23rd October and the supervisor informs us, then the topic will be assigned to you. The remaining topics will be available for assignment based on a lottery system in the Kick-off meeting for the seminar on 6th November, 2020. Only students with assigned topics will be accepted to our seminar.

List of Topics:

1) Load Balancing Techniques for Improving the Performance of 5G Core Networks (Assigned, Not Available!)
Supervisor: Endri Goshi
Description: The architecture of the Mobile Core Network in 5G and beyond is fundamentally different from its predecessors. Using Network Function Virtualization (NFV), the 5G Core Network Functions (NFs) are envisioned to be deployed in a Cloud-Native environment, following the example of IT services. This leads to a distributed Core Network, where new NF replicas are scheduled on-demand and they are placed in optimal locations to fulfill the SLAs. However, having more replicas does not necessarily solve the problem of overloading the Core Network, thus it is necessary that load balancing techniques are implemented in order to increase the overall performance of the Core Network as well as achieving the desired QoE. The goal of this topic is to investigate the current mechanisms that are used to load balance the control traffic that goes through the Core Network and provide a clear analysis of the advantages and drawbacks of these solutions.

References:
[1] T. V. Kiran Buyakar, H. Agarwal, B. R. Tamma and A. A. Franklin, "Prototyping and Load Balancing the Service Based Architecture of 5G Core Using NFV," 2019 IEEE Conference on Network Softwarization (NetSoft), Paris, France, 2019, pp. 228-232, doi: 10.1109/NETSOFT.2019.8806648.
[2] V. Nguyen, K. Grinnemo, J. Taheri and A. Brunstrom, "On Load Balancing for a Virtual and Distributed MME in the 5G Core," 2018 IEEE 29th Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), Bologna, 2018, pp. 1-7, doi: 10.1109/PIMRC.2018.8580693.
[3] B. Ma, B. Yang, Y. Zhu and J. Zhang, "Context-Aware Proactive 5G Load Balancing and Optimization for Urban Areas," in IEEE Access, vol. 8, pp. 8405-8417, 2020, doi: 10.1109/ACCESS.2020.2964562.

2) Reinforcement learning for Resource Allocation in LiFi-RF HetNets with Multi-Homing Users (Assigned, Not Available!)
Supervisor: Hansini Vijayaraghavan
Description: LiFi is becoming an increasingly important technology as the unlicensed ISM band gets overcrowded. The integration of LiFi into existing communication systems is best envisioned in the form of wireless Heterogeneous Networks or HetNets where the short-range, additional capacity providing LiFi cells complement the broader coverage providing RF cells. In order to fully utilize the potential of an intertechnology interference free LiFi-RF HetNet, a user device must be capable of using multiple network interfaces simultaneously. Thanks to Multi path solutions like MPTCP this is possible. The challenge in resource allocation in a MPTCP-enabled heterogeneous network lies in deciding a policy to schedule data packets onto the multiple paths with heterogeneous characteristics (eg. Delay, loss). The dynamic nature of these paths due to their wireless nature is an added challenge to the scheduling problem. Reinforcement learning has emerged as a powerful tool in wireless resource allocation and can be a promising solution to the packet scheduling problem. This work involves carrying out a survey of the different papers on using reinforcement learning for packet scheduling in MPTCP and comparing the different algorithms in terms of their suitability for a LiFi-RF Heterogeneous Network.

References:
[1] H. Zhang, W. Li, S. Gao, X. Wang and B. Ye, "ReLeS: A Neural Adaptive Multipath Scheduler based on Deep Reinforcement Learning," IEEE INFOCOM 2019 - IEEE Conference on Computer Communications, Paris, France, 2019, pp. 1648-1656, doi: 10.1109/INFOCOM.2019.8737649
[2] H. Wu, Ö. Alay, A. Brunstrom, S. Ferlin and G. Caso, "Peekaboo: Learning-Based Multipath Scheduling for Dynamic Heterogeneous Environments," in IEEE Journal on Selected Areas in Communications, vol. 38, no. 10, pp. 2295-2310, Oct. 2020, doi: 10.1109/JSAC.2020.3000365.

3) Optimal 5G functional split selection based on resource consumption (Assigned, Not Available!)
Supervisor: Alberto Martinez Alba
Description: The 5G base stations (gNBs) are divided into centralized and distributed units, due to the advantages of centralized hosting of the mobile functions. One of this advantages is cost reduction, since centralized units may be deployed in large data centers, whose operational cost is substantially lower than those of small distributed units. Owing to this, recent research has proposed to dynamically change the functional split with the objective of minimizing the deployment and operational costs depending on the instantaneous traffic load. In this topic, the student will investigate the benefits and weaknesses of this approach.

References:
[1] Matoussi, Salma, et al. "5G RAN: Functional Split Orchestration Optimization." IEEE Journal on Selected Areas in Communications 38.7 (2020): 1448-1463.

4) A survey on fingerprinting techniques (Available)
Supervisor: Patrick Kalmbach
Description: Fingerprinting of web-sites, applications, clients and traffic using statistical patterns has become a vital part of network management. Your task is to perform a literature research on the topic. Work of interest includes website-fingerprinting, TLS fingerprinting and statistical traffic classification.

References:
[1] H. Kim, K. Claffy, M. Fomenkov, D. Barman, M. Faloutsos, and K. Lee, “Internet Traffic Classification Demystified: Myths, Caveats, and the Best Practices,” New York, NY, USA, 2008, doi: 10.1145/1544012.1544023.
[2] T. Wang, X. Cai, R. Nithyanand, R. Johnson, and I. Goldberg, “Effective Attacks and Provable Defenses for Website Fingerprinting,” in Proceedings of the 23rd USENIX Conference on Security Symposium, USA, 2014, pp. 143–157.
[3] V. F. Taylor, R. Spolaor, M. Conti, and I. Martinovic, “Robust Smartphone App Identification via Encrypted Network Traffic Analysis,” IEEE Transactions on Information Forensics and Security, vol. 13, no. 1, pp. 63–78, Jan. 2018, doi: 10.1109/TIFS.2017.2737970.
[4] C. López Romera, “DNS Over HTTPS Traffic Analysis and Detection,” Master Thesis, Universitat Oberta de Catalunya (UOC), Oberta, Spain, 2020.
[5] H. F. Alan and J. Kaur, “Client Diversity Factor in HTTPS Webpage Fingerprinting,” in Proceedings of the Ninth ACM Conference on Data and Application Security and Privacy, New York, NY, USA, 2019, pp. 279–290, doi: 10.1145/3292006.3300045.

5) Reinforcement Learning approaches for efficient Radio Access Network Slicing (Assigned, Not Available!)
Supervisor: Arled Papa
Description: Network slicing is envisioned in 5G era as one of the main techniques to achieve the deployment of heterogeneous applications in a common physical infrastructure. While network slicing promises cost reduction and flexibility in next generation networks, it also poses significant challenges especially in terms of slice isolation. In particular, the problem of slice isolation becomes extremely challenging in the Radio Access Networks (RANs) due to the dynamic nature of wireless environments and tight coupling of the wireless channels with the scarce wireless resources. In that regard efficient techniques are required for an adequate resource allocation in a timely manner. Such resource allocation problems should often take into account various parameters such as wireless channel updates, user mobility and traffic predictions which are often not available and known in advance. To this end machine learning techniques and especially reinforcement learning is foreseen as a potential solution to learn/predict these parameters and therefore perform a better resource allocation in real time. The aim of this seminar for the student will be to identify the main state of the art papers with respect to network slicing using reinforcement learning and perform a survey comparing them.

References:
[1] Sciancalepore, Vincenzo, Xavier Costa-Perez, and Albert Banchs. "RL-NSB: Reinforcement learning-based 5G network slice broker." IEEE/ACM Transactions on Networking 27.4 (2019): 1543- 1557.
[2] Qi, Chen, et al. "Deep reinforcement learning with discrete normalized advantage functions for resource management in network slicing." IEEE Communications Letters 23.8 (2019): 1337-1341.

6) Achiving high-utilization in Software-Defined WAN (Assigned, Not Available!)
Supervisor: Nemanja Deric
Description: Achieving high utilization in Software-Defined wide area networks (WAN) is a challenging problem. To overcome this problem, Google proposed and deployed their own novel solution (i.e., B4) in their private WAN. However, there are also a few other approaches targeting the same problem (e.g., SWAN developed by Microsoft). In this seminar, the goal is to summarize B4 and to provide a comprehensive overview of the other state-of-the-art approaches.

References:
[1] Hong, Chi-Yao, et al. "B4 and after: managing hierarchy, partitioning, and asymmetry for availability and scale in google's software-defined WAN." Proceedings of the 2018 Conference of the ACM Special Interest Group on Data Communication. 2018.