Student Work

Hyper-connected meshes are a growing idea that improves the connectivity in core networks. This approach adds several links to the existing core network to make it more dense and increases the number of simple paths in the network. There are several approaches to add these links to the network. In this work, we compare the robustness of the different approaches on different topologies and identify the best approach to add links in terms of robustness. 

We already have our tool to develop robustness surfaces for the given network. The student is expected to use this tool and compare the robustness of different approaches and make meaningful inferences.

[1] Manzano, M., Sahneh, F., Scoglio, C., Calle, E. and Marzo, J.L., 2014. Robustness surfaces of complex networks. Scientific reports, 4(1), p.6133.

[2] Rueda, D.F., Calle, E. and Marzo, J.L., 2017. Robustness comparison of 15 real telecommunication networks: Structural and centrality measurements. Journal of Network and Systems Management, 25(2), pp.269-289.

This work is in collaboration with Christofer Vásquez from the University of the Bundeswehr, Munich.

Prerequisites

How can the capacity of a quantum network nodes be assigned optimally to the connection of the node, such that the rejection rate and wasted resources is minimized.

How do different routing metrics for routing in quantum networks impact rejection rates and wasted resources.

In memory constraint quantum networks, how does entanglement need to be reallocated, such that most of the entanglement requests can be fullfilled.

The interdependence of communication networks and electrical power grids is a topic of discussion in recent times. Though researchers try to map the interdependence, there are no topology-level frameworks that have data on both communication networks and electrical power grids. This work aims to generate such topology-level frameworks to publish as open-source to encourage research along this direction.

Prerequisites

A cut set of a flow is a set of components when removed, cause the flow to fail. A cut set is minimal if it cannot be further reduced. Based on minimal cut sets for flows, this work aims to investigate a potential centrality metric to determine the importance of particular nodes in the network. 

Prerequisites

The goal is to optimize the purification for a given entanglement generation method.

The goal is to see whether the entanglement should be swapped at the request phase or in the ack phase, to increase the resulting fidelity.

The transition from 4G to 5G introduced a highly service-based and cloud-native core network architecture. While the 5G User Plane Function (UPF) provides efficient packet forwarding, security functionality traditionally associated with packet gateways in previous generations has largely been separated from the user plane. To address security concerns, the 3GPP architecture introduces the Inter PLMN User Plane Security (IPUPS) function, which can be deployed alongside the UPF and acts as a non-transparent firewalling entity.

However, recent developments indicate that transparent firewalling approaches may also be feasible in 5G environments, potentially offering advantages in terms of performance, deployment flexibility, and operational simplicity. At the same time, the increasing use of IPsec-based protection for inter-operator roaming interfaces raises new challenges and opportunities for different firewall architectures.

This thesis aims to investigate and compare transparent and non-transparent security approaches in 5G core networks through practical implementation and experimental evaluation.

Motivation

Emerging 5G applications, particularly in Ultra-Reliable Low-Latency Communications (URLLC), require strict guarantees on latency and reliability. Hybrid Automatic Repeat Request (HARQ) is a key mechanism for improving transmission reliability, but retransmissions, decoding delays, and feedback latencies introduce variability in packet delays. While analytical models provide insights into the resulting Delay Violation Probability (DVP), their accuracy may be limited under practical deployment conditions. Therefore, data-driven approaches based on real measurements are needed to improve delay prediction and support latency-aware network optimization.

Problem Statement

Existing delay analysis frameworks for 5G HARQ primarily rely on analytical modeling assumptions and are often validated only through simulations. Furthermore, current machine learning models for DVP prediction do not fully exploit real-world testbed data and typically lack support for downlink-specific HARQ characteristics. As a result, there is a gap between theoretical delay analysis and practical delay prediction in operational 5G systems.

Thesis Goal

The goal of this thesis is to collect HARQ-related delay measurements from a 5G testbed and use them to validate and extend an existing machine learning model for predicting Delay Violation Probability. The work will investigate the relationship between HARQ retransmissions, decoding latency, feedback latency, and overall packet delay. In addition, the existing prediction model will be extended to incorporate downlink HARQ features, and its prediction accuracy will be evaluated using experimentally collected data.

5G networks increasingly serve mission-critical applications such as industrial automation, autonomous vehicles, and remote control. These use cases require highly deterministic communication with strict bounds on latency, jitter, and reliability. Time-Sensitive Networking (TSN), originally designed for wired industrial Ethernet, provides mechanisms for deterministic communication and is now being integrated with 5G to provide mobility. 

Although the reliability mechanism, Frame Replication and Elimination for Reliability (FRER), is supported in TSN-5G systems, the configuration is typically fixed or manually configured. Real 5G environments, however, exhibit dynamic and sometimes unpredictable wireless channel conditions. This necessitates a Central Network Controller that dynamically reconfigures the reliability mechanisms. 

This thesis aims to simulate and analyze the dynamic management of FRER and evaluate their performance under varying channel conditions.

Thesis Outline

The increasing use of cellular-connected drones (UAVs), especially for beyond-visual-line-of-sight applications, introduces significant challenges for airspace security, critical infrastructure protection, and regulatory compliance. Unlike traditional drones controlled via short-range radio, modern UAVs increasingly rely on 5G/6G networks for command, control, and video transmission. This shift opens up a novel opportunity: mobile network infrastructure itself can be leveraged as a large-scale sensor system for drone detection. The objective of this thesis is to investigate whether cellular-connected drones can be detected using mobile network data.

Work Items

The thesis is structured into several closely connected stages:

• comprehensive literature review that systematically surveys existing approaches to drone detection in mobile networks
• in-depth analysis of RAN and core network parameters at the MAC layer and above to identify metrics that may carry discriminative information
• generation and collection of representative data
• investigate different drone detection concepts based on the identified metrics

Depending on the type of thesis and the progress during the work, the scope may be adjusted, and not all work items need to be completed. For more details regarding the individual work items, do not hesitate to contact us.

Prerequisites

In this thesis, intent-based prediction of the MAF execution using ML is explored.

In this thesis, the holistic placement of MAFs using ML for placement and scheduling is explored.

• Optimizing Dual Connectivity
• Robust Optimization Problem
• RL solution

Traditional mobile networks are typically based on a single access technology or single radio link connections, which makes them prone to link failures, congestion, and interference. These limitations result in service interruptions, high latency, and reduced throughput, which can compromise safety and efficiency in mission-critical applications. 
This work involves developing and implementing wireless resource management algorithms by 

• formulating mathematical optimization problems
• solving the problems with heuristics or machine learning-based approaches
• evaluating the algorithms in simulators or hardware testbeds

If you are interested in this work, please send me an email with a short introduction of yourself and your interests along with your CV and grade transcript.

Note: There are multiple thesis/internship positions available within this umbrella topic. The specific topic could be adapted to suit your interests.

Prerequisites

The idea is to implement attacks against the 5G SEPP and implement countermeasures into the Open5GS SEPP implementation. The effect of the attacks and countermeasures shall be evaluated.

This will be done on the already existing 5G Roaming Testbed at LKN, which is dockerized and can run on any two machines (VMs, PCs, ..).

Due to the increasing demands of Industry 4.0, 5G networks are increasingly used to support mobility in Time-Sensitive Networking (TSN) applications such as remote robot control and IoT systems. These applications must support a wide variety of traffic types, including best-effort, rate-sensitive, and low-latency traffic streams.

Recent research [1] proposes joint scheduling between 5G and TSN, focusing on an aligned resource allocation of the 5G RAN scheduler and TSN mechanisms Time-Aware Shaper (TAS) and Per-Stream Filtering and Policing (PSFP). In particular, deterministic scheduling in the uplink is sensitive to jitter because of the grant-based control signaling mechanism in 5G. To address this challenge, [1] proposes a heterogeneous uplink scheduler that combines grant-based and grant-free transmissions. This approach enables static allocation of radio resources for time-sensitive traffic streams while still allowing dynamic allocation for traffic with non-deterministic sending behavior.

However, the approach has so far been evaluated only through simulations, while experimental validation in real testbed environments remains an open challenge. The goal of this thesis is therefore to experimentally evaluate the TSN-5G scheduling framework in an end-to-end testbed. The work involves the following tasks:

• Implement heterogenous 5G RAN Scheduler in OpenAirInterface RAN and as xApp
• Configure TSN Switches (TAS, PSFP) considering the 5G uplink bridge delay
• Experimental evaluation of the scheduling framework under different scenarios, including
• different TDD patterns
• different traffic scenarios

[1] L. Becker, Y. Deshpande, and W. Kellerer, “Joint Resource Allocation to Transparently Integrate 5G TDD Uplink with Time-Aware TSN,” arXiv preprint arXiv:2511.23373, 2025.

Prerequisites

Motivation
Autonomous drone swarms represent a key research topic in modern robotics and communication systems. They are increasingly applied in safety-critical domains such as disaster response, industrial inspection, and military reconnaissance. A fundamental prerequisite for the reliable operation of such systems is a robust, real-time capable, and secure communication between the individual drones.

In practice, many existing systems rely on centralized control or coordination entities, which introduce single points of failure. Furthermore, distributed swarms often assume an implicit trust model, where received messages are processed without cryptographically secured authorization or consistency mechanisms. This makes the overall system vulnerable to node failures, communication faults, or targeted attacks.

The Data Distribution Service (DDS) has established itself as a powerful standard for decentralized, real-time publish/subscribe communication and is widely used in the context of ROS 2. However, DDS primarily addresses latency, reliability, and quality-of-service requirements, while it does not provide mechanisms for trust establishment, identity management, or consistent decision-making in distributed systems.

Blockchain technology offers decentralized consensus mechanisms, immutable state records, and cryptographic identities, providing promising solutions to these challenges. Due to its limited real-time capabilities, however, blockchain technology is not suitable for the direct exchange of time-critical sensor data.

The combination of DDS and blockchain technology therefore enables the realization of a hybrid communication model that combines real-time data exchange with decentralized trust and fault tolerance.

Objective of the Thesis
The objective of this master’s thesis is the development and evaluation of a decentralized and trusted communication architecture for autonomous drone swarms that:
• avoids centralized control entities,
• enables real-time data transmission,
• and provides cryptographic protection for safety-critical decisions.

This leads to the following research question:
How can a hybrid communication architecture based on DDS and blockchain technology be designed to meet the real-time requirements of autonomous drone swarms while providing a decentralized trust and authorization structure?

The 5G Service-Based Architecture (SBA) introduces new challenges and opportunities in the design of roaming interfaces between mobile network operators. Central to secure inter-operator communication are the Security Edge Protection Proxy (SEPP) and IP eXchange (IPX) network provider entities, which facilitate secure signaling across trust boundaries.

An existing Python-based simulation framework models a fixed 5G roaming topology with four entities: vSEPP, vIPX, hIPX, and hSEPP. While this setup is suitable for basic testing, it lacks scalability and flexibility for simulating realistic, multi-operator roaming scenarios.

This thesis aims to refactor and extend the SEPP and IPX implementations to support scalable, multi-connection environments, enabling dynamic routing and connection management across a larger simulated network. The SEPP must unify the roles of vSEPP and hSEPP, while the IPX must support parallel connections and enforce routing rules based on operator policies and network topology. To achieve realistic routing behavior, the thesis will incorporate the FRRouting (FRR) suite, an open-source routing stack supporting protocols such as BGP, OSPF, and IS-IS. FRR will be used to simulate routing decisions and path selection within the IPX network, enabling policy-based and topology-aware message forwarding.

Objectives

1. Unified SEPP Design: Develop a single SEPP module that combines both vSEPP and hSEPP functionalities, capable of handling multiple concurrent connections and maintaining session state.
2. Enhanced IPX Module: Extend the IPX implementation to support: Multiple parallel connections. Routing logic based on operator identifiers, trust relationships, and message types. Integrate with the FRRouting suite to simulate realistic routing behavior.
3. Scalability Framework: Design a flexible configuration system to instantiate multiple SEPP and IPX entities dynamically, simulating a larger roaming network.
4. Testing & Validation: Create test scenarios to validate connection handling, routing correctness, and protocol compliance.
5. Documentation & Thesis Writing: Produce comprehensive documentation and a formal thesis de-

tailing design decisions, implementation, and evaluation results.

Prerequisites

Goal: To obtain a rule-based heuristic from a well-performing AI agent for an NP-hard optimization problem. 

AI models, especially deep learning systems, often operate as "black boxes," making complex decisions without revealing how they arrived at a specific conclusion. Explainable AI provides insights into the decision-making process, fostering transparency. As future 5G/6G networks evolve, mobile operators will increasingly rely on AI-driven solutions. For widespread adoption, it's crucial that operators trust these systems, a trust that is far more likely to be built if they can clearly understand how the AI arrives at its decisions.

First, we will focus on extracting the underlying rules that guide AI decision-making. Next, we will aim to develop a rule-based heuristic that can replicate the AI’s performance, offering a simpler yet equally effective solution.

Prerequisites

Through its nationwide communication infrastructure, Deutsche Telekom operates a large variety of services targeted at the needs of customers and their devices. With technological advances reaching many industries, the set of such networked daily-use devices includes not only phones but TV attachments and many more. Naturally, this combination of a high number of users plus the variety of services and devices produces a large amount of heterogeneous data. Unexpected events and anomalous behavior can easily cause service disruptions and even downtime for the system.Therefore, it is important to identify points within the streaming data that indicate deviations from normal system operation. In this context, the thesis aims to evaluate the ability to flag such anomalies early on or even predict them in advance, essentially creating an early warning model (EWM).

Prerequisites

In today’s business landscape, the demand for highly available, secure, and scalable networks is continuously increasing, particularly for large enterprises.

Conventional network management faces challenges such as complexity, with manual configurations being error-prone and time-consuming. It also struggles with scalability issues due to slow adaptation to changing needs and limited automation, which requires deep expertise. Modern solutions like SDN, NFV, and AI-driven automation address these problems by enabling dynamic, scalable, and policy-driven network management.

The traditional network management approach relies on manual implementation, requiring expertise in routing, Quality of Service (QoS), and encryption mechanisms. This results in high operational costs and makes the network prone to misconfigurations. Intent-Based Network Configuration Management is a modern approach to managing and automating networks, where the operator defines "what they want the network to do" (the intent) rather than specifying "how to configure the network" (manual steps). The system interprets these high-level intents and automates the necessary configurations and adjustments to achieve the desired outcome.

Prerequisites

This master’s thesis addresses the challenge of classifying encrypted traffic in large-scale networks. Traditionally, network operators could monitor and deeply analyze the traffic flowing through their infrastructure to optimize traffic management. However, since the rise of encrypted communications and security concerns, the details accessible by network operators have shrunk to a very high degree. Thus, the current project aims to propose and implement a method for in-network, flow-based encrypted traffic classification.

Alongside the increasing difficulty in classifying and analyzing encrypted traffic comes also the limitations of in-network classification in traditional networks. Traditional network monitoring and analysis methods suffer from little scalability. Therefore, the question addressed in this thesis is whether software-defined networks (SDN) are suitable for in-network traffic classification, relying on SDN-enabled switches distributed across the network operator’s infrastructure. For this classification task, a decision tree classifier implemented using the P4 language and running on the switches is proposed. Finally, the performance and accuracy of implementing such processing will be assessed.

Key points:

• Large-scale traffic classification

• ETC is a hot topic. It provides information to the network operator for optimizing traffic management

• SDN is the new networking paradigm that helps by relying on switches across the network for classification instead of relying on a central network element for it, which could be a bottleneck and could hinder getting real-time analytics.

• Decision Trees as the Machine Learning model used for traffic classification, based on previous work done in the field.

Prerequisites

The increasing complexity of video analysis tasks has led to the development of sophisticated machine learning applications that rely on multiple interconnected deep-learning models. However, deploying these applications on edge servers with limited resources poses significant challenges in terms of balancing response time, accuracy, and resource utilization.

This research aims to investigate the trade-offs between monolithic and microservice-based architectures for multi-model video analytics applications. We will leverage Apache Ray and Kubernetes to develop a comprehensive benchmarking and monitoring pipeline that systematically analyzes scaling configurations and resource utilization. Our goal is to identify strategies for reducing latency and communication overhead while handling the complexity of multi-model architectures.

Through case studies, we will implement two applications consisting of multiple ML models and explore various deployment configurations, from monolithic to fully microservice-based. We will analyze their performance under different scaling strategies and investigate the impact of model disaggregation on performance metrics during scaling.

The research will focus on addressing key challenges, including:

• Implementing effective monitoring and profiling across distributed services
• Balancing performance and resource utilization across multiple models
• Researching scalability solutions to meet strict latency requirements in edge computing scenarios

Our expected outcomes include a deeper understanding of the trade-offs between monolithic and microservice-based architectures and the development of strategies for optimizing the deployment of multi-model ML applications.

This project aims to identify, analyze, and mitigate antipatterns in AI applications that use microservices-based architectures in distributed systems. The focus will be on Apache Ray within a Kubernetes environment. The goal is to develop strategies to optimize performance and resource efficiency by addressing common antipatterns that can degrade system performance, such as inefficient communication between services, over-provisioning of resources, and poor task distribution.

To achieve this, the project will benchmark AI applications, including video surveillance and action detection systems, to understand how antipatterns emerge and affect key performance metrics like latency, throughput, and resource utilization.

The growing interdependence between optical access networks and electrical distribution systems is a key factor in maintaining the reliability of essential services in urban environments. As cities increasingly rely on both power and digital connectivity, the resilience of these interconnected infrastructures has become more critical than ever. This thesis explores the integration of various protection schemes within Passive Optical Networks (PON) and assesses their dependability on electrical distribution networks under multiple failure scenarios.

The protection schemes examined include the Unprotected scheme, as well as more robust configurations such as Type A (Feeder Fiber Protection), Type B (Dual Parented), Type C, and Type C2, which offer varying levels of reliability. These protection strategies aim to mitigate the risks posed by network failures, ensuring uninterrupted service in urban areas where dependability is essential.

A primary goal of this thesis is to develop a generic tool that can assess the dependability of optical and electrical networks, particularly within the context of FttX (Fiber to the X) architectures. While the focus will primarily be on FttC (Fiber to the Cabinet), the tool might also explore potential applications in FttB (Fiber to the Building) and FttH (Fiber to the Home) scenarios. By using OpenStreetMap (OSM) data to model real-world urban layouts, the tool will integrate these protection schemes into a PON planning system to design both optical and electrical network topologies.

Additionally, the tool will feature a failure analysis component that simulates different failure types – such as node, edge, and disaster failures – to generate robustness metrics. These analyses will be complemented by visualizations such as robustness surfaces and statistical breakdowns, offering insights into the resilience of urban infrastructures. Through this approach, the thesis aims to enhance the overall understanding of how optical and electrical networks respond to stress, with a focus on ensuring robustness and reliability in densely populated urban settings.

This thesis aims to achieve its objectives through a combination of Python scripting, data analysis libraries, and geospatial data sources. As a geospatial data source, OSM is planned to be used, and for processing and visualizing the networks Python will be used. Key Python libraries include Folium, which is employed to plot the networks interactively on maps, and NetworkX, used for creating and analyzing the structure of complex network topologies. Pandas can be used in managing and analyzing the large datasets generated from OSM and network simulations. Visualization tools such as Matplotlib, Seaborn and Folium will be utilized to create both static and interactive visual representations of the data, including network layouts and failure impacts. Together, these tools will enable a robust analysis of network dependability across optical and electrical infrastructures.

Prerequisites

The widespread deployment of Fiber-to-the-Home (FttH) in various markets has given consumers access to high-speed data connections over the extensive optical fiber-based core networks. End users usually access these networks with wireless devices using the 802.11 standard. However, as the next generation of IoT devices with VR and AI applications enter the market, the demand for improved bandwidth and latency has increased drastically. This has led to the development of Fiber-to-the-Room (FttR) networks, which seek to extend the fiber connections closer to the terminal devices by eliminating the bottleneck caused by traditional ethernet and wireless access. Current implementations utilize a cascaded XGPON optical distribution network (ODN) in convergence with WiFi 6. The ONU at the user’s premises is replaced with a Main FttR Unit (MFU), which acts as the OLT for the Sub FttR Units (SFU) in each room using point-to-multipoint architecture. The SFUs convert the optical signals to baseband and retransmit the data as WiFi frames.

These solutions keep both MAC and PHY layers between the optical transport and WiFi completely separated without using any synergies. This introduces various limitations to the FttR setup, such as degraded data rates, jitter / latency, and high power consumption. To solve these problems, a centralized wireless - optical access network (C-WAN) has been conceptualized, where the MFU becomes a central controller to enable seamless roaming, with deterministic low latency transmission and gigabit coverage for the entire room. Within the scope of this solution, a cross-domain MAC layer between optical and WiFi has been proposed to implement dynamic resource management and scheduling. Such a centralized and converged MAC protocol must reduce overhead and optimize the framing of both physical layers to minimize the processing delay at the SFUs. This thesis aims to design and evaluate such a minimal optical access protocol with a centralized controller and test its performance in an FttR setup. The ns-3 network simulator is used within the context of this thesis to implement, simulate, and analyze the designed protocol with C++. Extensive modules that simulate both XGPON and WiFi 6 are already implemented in ns-3 and can be modified as required.

Problem: BGP is performance-agnostic

Solution: incorporate a delay-related metric into the best-path selection process.

Approach: Estimate the round-trip prop_delay to destinations (/24s) within the routing table of the IXP

Goal: Evaluate if it is possible to outperform BGP’s route selection criterion, in terms of latency, with a measurement-based approach.

Providing Quality of Service (QoS) to emerging time-sensitive applications such as factory automation, telesurgery, and VR/AR applications is a challenging task [1]. Time Sensitive Networks (TSN) [2] and Deterministic Networks [3] were developed for such applications to guarantee ultra low latency, bounded latency and jitter, and zero congestion loss. The objective of this work is to develop a methodology to guarantee bounded End-to-End (E2E) latency and jitter in large-scale networks.

Prerequisites

In this IDP, we evaluate the potential of programmable dataplanes for medical use cases.

The research internship focuses on the robustness of drone communication links at the physical layer, especially with regard to jamming and spoofing attacks.

Possible tasks include:

• Literature research on PHY-layer robustness, jamming, and spoofing
• Identification of relevant attack scenarios and possible countermeasures
• Support with the implementation of countermeasures in MATLAB and Python
• Validation of the approaches using computer simulations
• Support with the implementation of SDR-based setups in GNU Radio
• Support with the setup of hardware demonstrators, integration on drone platforms, and execution of lab tests and flight experiments

Motivation
Time-Sensitive Networking (TSN) enables deterministic communication in Ethernet networks by providing mechanisms for bounded latency, low jitter, and traffic isolation. One important TSN mechanism is the Time-Aware Shaper (TAS), defined in IEEE 802.1Qbv, which controls the transmission of different traffic classes by opening and closing queue gates according to a cyclic schedule. This allows time-critical traffic to be transmitted in reserved, isolated time slots.

In industrial communication systems, 5G is increasingly considered as a flexible wireless extension of TSN. However, unlike wired TSN links, a 5G link introduces variable delay due to wireless channel conditions, scheduling, and retransmissions. This delay variation, or jitter, can disturb deterministic end-to-end communication.

A promising approach is to use TAS scheduling in the wired TSN domain to compensate for the measured delay behavior of the 5G link. To do this effectively, the 5G link must first be characterized experimentally. Based on measured delay and jitter distributions, suitable TAS schedule parameters can be derived, such as gate opening times, guard intervals, cycle times, and offsets. 

Problem Description

TAS schedules are typically designed for deterministic wired networks where link delays are relatively stable and predictable. In a 5G-TSN scenario, however, the 5G segment behaves as a stochastic link with variable latency. If the TAS schedule is configured without considering this delay variation, packets may arrive too early or too late for the intended transmission window. This can lead to additional queuing delay, missed transmission slots, increased jitter, or loss of deterministic behavior.

The key problem is therefore to determine how to configure TAS parameters when a 5G link is part of the end-to-end communication path. This requires understanding the delay behavior of the 5G link under different operating conditions and deriving practical rules for TAS schedule configuration from this characterization.

The internship will focus on a testbed-based investigation. The 5G link will be measured experimentally, the resulting delay and jitter behavior will be analyzed statistically, and initial TAS scheduling guidelines will be derived. If time permits, the work may also formulate the TAS configuration as an optimization problem and evaluate optimized schedules.

Goal

The main goal of this internship is to characterize the delay behavior of a 5G link in a TSN-related testbed and to derive practical guidelines for configuring TAS schedules to reduce end-to-end jitter.

The minimum expected outcome is:

1. A completed measurement campaign of the 5G link delay and jitter behavior.

2. A statistical characterization of the measured 5G delay distribution.

3. A first set of TAS configuration guidelines based on the measured delay behavior.

4. If feasible, an initial testbed evaluation of the proposed TAS settings.

An optional extended goal is to formulate the TAS schedule configuration as an optimization problem, for example, minimizing jitter or end-to-end latency while respecting measured delay bounds and TAS cycle constraints.

• TN-NTN simulator

• ATSSS simulator

Serverless computing is a cloud computing paradigm that separates infrastructure management from software development and deployment. It offers advantages such as low development overhead, fine-grained unmanaged autoscaling, and reduced customer billing. From the cloud provider's perspective, serverless reduces operational costs through multi-tenant resource multiplexing and infrastructure heterogeneity.

    However, the serverless paradigm also comes with its challenges. First, a systematic methodology is needed to assess the performance of heterogeneous open-source serverless solutions. To our knowledge, existing surveys need a thorough comparison between these frameworks. Second, there are inherent challenges associated with the serverless architecture, specifically due to its short-lived and stateless nature.

The efficient allocation of communication resources is a key challenge in future 6G in-X subnetworks, particularly in highly dynamic environments such as urban vehicular scenarios. In this work, a demonstrator for resource allocation schemes in 6G in-X subnetworks will be developed. The student will use the Simulation of Urban MObility (SUMO) framework to model realistic vehicular mobility and generate network scenarios. A backend will be implemented to integrate and execute different resource allocation algorithms, while a frontend will provide interactive visualization of network conditions and algorithm performance. The resulting demonstrator will enable the analysis and comparison of resource allocation strategies, highlighting their impact on network efficiency and service performance in dynamic vehicular environments.

Modern mobile robots often use cameras to perceive their environment or to allow remote control by a human operator. These video streams require a reliable network connection, especially when transmitted over 5G. However, network conditions in 5G can change dynamically due to movement, signal quality, or network load. This can lead to delays, reduced video quality, or even connection issues. This project focuses on making the robot “smart” in how it sends video data: instead of using fixed settings, the video stream should automatically adapt to current network conditions.

The aim is to develop an application running on a robot’s onboard computer that adjusts the quality of a stereo camera video stream based on real-time 5G network conditions. The system should balance video quality and reliability, i.e., video stability and availability, to achieve smooth and efficient transmission.

Clustering is already used in our cities to distribute splitters for managing network operations. However, the current distribution does not effectively reduce the costs caused by huge urban area with long fiber lengths. To solve this, we can use k-means clustering to group the splitters and optimize how fibers are laid out in the city.

With k-means clustering, we group users (ONUs) in an area to create a virtual center. This center is chosen so that the total distance to all users in the area is minimized. We then find a nearby point where a splitter can be placed, ensuring the distance from the splitter to the users is as short as possible.

Next, we cluster the remote nodes (RNs) where splitters are located. By grouping nearby splitters, we can combine their connections into one line to the central office (OLT). This line will be organized by a new splitter near to the center of the old splitter's cluster center. This helps reduce the total fiber length needed and lowers deployment costs.

Prerequisites

...

The Internship aims to analyze Open Source Optical Network Data for performance prediction of Optical Networks and develop data-driven methods for quality of transmission estimation.

The primary objective of this work is to help maintain and extend the LKN 5G Roaming Testbed. The testbed is now available as open-source software on GitHub and it is also part of the student position to keep that repository up to date: https://github.com/tum-lkn/LKN_5G_Roaming_Testbed

The main use of the LKN 5G Roaming Testbed is to perform security investigations, but also present the testbed in various demonstrations. While security investigations in more detail are part of other student works like research internships and master theses, helping to prepare for the demos is also part of this working student position.

Connected robots need fast and reliable communication, but relying on a single wireless link can lead to failures during critical tasks. This student job focuses on combining technologies such as 6G-cellular, WiFi, and LiFi into a unified 6G network, improving reliability, reducing latency, and increasing data rates for next-generation robotic systems.

The goal of this work is to help implement and maintain a 6G multi-technology testbed and then use the testbed to investigate the performance of multipath networks. In this work, you have the opportunity to work with open source software mobile networks software as well as learn about a new technology – LiFi. This working student position can also be followed up with a research internship or thesis.

If you are interested in this work, please send me an email with a short introduction of yourself along with your CV and grade transcript.

Prerequisites

Reliability Block Diagrams (RBDs) are visual aids that help evaluate the availability of a system. In communication networks, the availability evaluation is often computation and memory-intensive owing to the large size of the topologies. Hence, we investigate a novel approach of evaluating RBDs using minimal path sets.

Prerequisites

Molecular communication (MC) is an alternative to classical electromagnetic wave-based communication, where molecules are used for information exchange. MC is expected to enable in-body networks for future medical applications in the Internet of Bio-Nano Things, a vision for 6G and beyond.

We seek a working student to help us operate and maintain a molecular communication networks testbed at the chair. The testbed is based on ink molecules transmitted through a water-filled tube system with a background flow. Spectral absorption measurements are used to detect the information molecules.

What you will do:

• Set up the tubes, pumps, sensors, and other components
• Program microcontrollers to control microfluidic pumps and spectral sensors
• Debugging and fault detection in both software and hardware
• Implement data collection and visualization methods
• Support with experiments conducted on the testbed

Prerequisites

Future medical applications put stringent requirements on the underlying communication networks in terms of highest availability, maximal throughput, minimal latency, etc. Thus, in the context of the 6G-life project, new networking concepts and solutions are being developed.

For the research of using 6G for medical applications, the communication and the medical side have joined forces: While researchers from the MITI group (Minimally invasive Interdisciplinary Therapeutical Intervention), located at the hospital "Rechts der Isar", focus on the requirements of the medical applications and collecting needed parameters of patients, it is the task of the researchers at LKN to optimize the network in order to satisfy the applications' demands. The goal of this joint research work is to have working testbeds for two medical testbeds located in the hospital to demonstrate the impact and benefits of future 6G networks and concepts for medical applications.

Your task during this work is to implement the communcation network for those testbeds. Based on an existing open-access 5G network implementation, you will implement changes according to the progress of the current research. The results of your work, working 6G medical testbeds, will enable researchers to validate their approaches with real-world measurements and allow to demonstrate future 6G concepts to research, industry and politics.

In this project, you will gain a deep insight into how communication networks, especially the Radio Access Network (RAN), work and how different aspects are implemented. Additionally, you will understand the current limitations and weaknesses as well as concepts for improvement. Also, you will get some insights into medical topics if interested. As in such a broad topic there are many open research questions, you additionally have the possibility to also write your thesis or complete an internship.

Prerequisites

Beginning possible from 1st October 2024

Future medical applications put stringent requirements on the underlying communication networks in terms of highest availability, maximal throughput, minimal latency, etc. Thus, in the context of the 6G-life project, new networking concepts and solutions are being developed.

For the research of using 6G for medical applications, the communication and the medical side have joined forces: While researchers from the MITI group (Minimally invasive Interdisciplinary Therapeutical Intervention), located at the hospital "Rechts der Isar", focus on the requirements of the medical applications and collecting needed parameters of patients, it is the task of the researchers at LKN to optimize the network in order to satisfy the applications' demands. The goal of this joint research work is to have working testbeds for two medical testbeds located in the hospital to demonstrate the impact and benefits of future 6G networks and concepts for medical applications.

Your task during this work is to implement the communcation network for those testbeds. Based on an existing open-access 5G network implementation, you will implement changes according to the progress of the current research. The results of your work, working 6G medical testbeds, will enable researchers to validate their approaches with real-world measurements and allow to demonstrate future 6G concepts to research, industry and politics.

In this project, you will gain a deep insight into how communication networks, especially the Radio Access Network (RAN), work and how different aspects are implemented. Additionally, you will understand the current limitations and weaknesses as well as concepts for improvement. Also, you will get some insights into medical topics if interested. As in such a broad topic there are many open research questions, you additionally have the possibility to also write your thesis or complete an internship.

Prerequisites

The Wireless Sensor Networks lab offers the opportunity to develop software solutions for the wireless sensor networking system, targeting innovative applications. For the next semester, a position is available to assist the participants in learning the programming environment and during the project development phase. The lab is planned to be held on-site every Tuesday 15:00 to 17:00.

Prerequisites