Student Work

Thanks to the ease of deployment and low cost, Wireless Sensor Networks (WSN) is a good candidate technology for many applications. In the context of industrial applications, WSN provides the necessary communication infrastructure for sensors and actuators to wirelessly operate in closed-loop control systems called Networked Control Systems (NCSs).

In state-of-the-art NCSs, a remote state estimator calculates the Minimum Mean Square Error (MMSE) estimate of the system’s state, which is then used by the controller to compute the actuation command. WSN devices, however, are subject to packet loss due to the low-power communication technology and external disturbances. This effect severely degrades the quality of estimation and can lead to the instability of the system under control. For this reason, an MMSE estimator operating over lossy WSN must take the stochastic packet loss into account [1,2].

In this work:

• An industrial wireless sensor network will be deployed using Zolertia Re-Motes and the Contiki-NG firmware.
• A measurement-based model of the WSN channel quality will be derived and included in the MMSE filter model.
• The quality of estimation of the MMSE filter will be evaluated and compared to [1].

This work will consist of the significant amount of programming for embedded devices, and for a large part also of processing the measurements data.

[1] Soleymani, Touraj, et al. "Covariance-based transmission power control for estimation over wireless sensor networks." 2018 European Control Conference (ECC). IEEE, 2018.

[2] Zoppi, Samuele, et al. "Transmission Power Control for Remote State Estimation in Industrial Wireless Sensor Networks." arXiv preprint arXiv:1907.07018 (2019).

• Strong C and Python programming skills.
• Good knowledge of wireless systems and protocols (WSN).
• Basic knowledge of LTI control systems. 
• Experience with python data processing tool is beneficial.
• Experience with embedded systems is beneficial but not required.

Live Virtual Machine (VM) migration is an unavoidable operation in virtualized data centers. In this process, a VM is moved from a server to another, with low service downtime. In this work, we plan to impement and evaluate the performance of of using multipath VM migration.

programming, computer networks, algorithms

amir.varasteh@tum.de

In this work, we plan to implement several Virtual Network Functions (VNFs) sucn as NAT, Firewall, IDS, etc. on Virtual Machines and Contrainers. For more info, please contact us.

programming, linux

amir.varasteh@tum.de

The goal in this work is to measure the performance overhead of virtual machine managment using different server hypervisors (KVM, Xen, VMWare, etc.). For more info, please contact us.

programming, linux, virtualization

amir.varasteh@tum.de

A Service Function Chain (SFC) can be deployed by utilizing Network Function Virtualization (NFV) concept. A challenge during the lifetime of SFCs can over-/underload of a Virtual Network Function (VNF). To solve this issue, firsly, the under-/overload bottleneck should be identified. Secondly, an algorithm must be proposed to solve the under-/overloaded situation efficiently.

In this work, we plan to design and evaluate an optimal strategy to identify and cope with under/overloaded situations in deployed SFCs. We will focus on a distributed optimization approach, called Alternating Direction Method of Multipliers (ADMM).

Mathematical Optimization, Algorithms, Java/Python

amir.varasteh@tum.de

Introduction to Software Defined Networking (SDN) and Network Virtualization paradigms has brought new challenges to tackle. In this environment, network switches communicate to SDN controllers through a hypervisor. Therefore, considering a network graph, three questions may come up: 1) Where to deploy SDN controllers? 2) Where to deploy SDN hypervirors? 3) How to connect network switches to the respective SDN controllers to meet the traffic dynamicity and QoS requirements?

In this work, we plan to design and evaluate an optimization framework to optimally answer the above questions.

Mathematical Optimization, Algorithms, Java / Python

amir.varasteh@tum.de

In this work, we plan to develop and implement an algorithm based on a metaheuristic e.g., Particle Swarm Optimization (PSO). This algorithm should be used to solve a multi-period VM placement and routing problem for multi-media services in airplanes. For more details, please send an email.

Experience in Python or Java, Optimization, Algorithms

amir.varasteh@tum.de

In this work, we would like to implement a nested approach to make three decisions in a Space-Air-Ground Integrated network: 1) Where to place VMs to provide Internet-based services for airplane passengers? 2) How to route the traffic from airplane to the VM? 3) Are VM migrations necessary over time? 

Experience in Python or Java, Mathematical Optimization, Algorithms

amir.varasteh@tum.de

Resource Management (VM Location, Routing, and Migration) in while airplanes are flying. An iterative approach tackles the problem by solving the small-sized problems iteratively. In this work, we would like to implement and evaluate this approach in a Space-Air-Ground Integrated scenario, where we provide QoS-guranteed Internet-based services to passengers while airplanes are flying.

Experience in Python or Java, Mathematical Optimization, Algorithms

amir.varasteh@tum.de

The lecture Broadband Communication Networks (Prof. Wolfgang Kellerer) teaches network-related methods of mobile communication: WIFI, 2G to 5G cellular networks, etc. In order to bridge the gap between the methods and real life, innovative teaching concepts shall be developed in form of short podcasts where the students learn in short episodes about wireless communication and networking in day to day scenarios. In the podcasts the student should also be introduced to short exercises they can perform on their own.

In order to support designing these podcasts Pro. Kellerer is looking for a student experienced in programming and maybe in podcasts to help him.

Very good programming skills; experience with podcasts; experience with app programming (on android/ios)

Access algortihms are a critical part of 5G networks. Currently, algorithm without direct feedback is used as access algorithms and these show unstable behavior. The main access algorithm with feedback, that is "Tree algorithms" are a prominent candidate to support reliable machine type communication in 5G. 

Another hot topic is the multi packet reception with orthogonalization techniques in code, power or frequency that is then decoded with successive interference cancellation techniques. This is simplified as K multipacket reception to investigate the capability for access algorithms.

In this thesis we plan to analytically investigate the behavior tree algorithms with K packet reception. The thesis will require:

• Good stochastic analysis knowledge
• Good coding in any language for simulation purposes
• Good understanding of the wireless communication
• Motivation to investigate mathematical analysis of system models.

The thesis will be co-supervised by Assoc. Prof. Cedomir Stefanovic from Aalborg University.

We offer a working student position to manage the LKN 5G testbed. The student will document and mantain the existing code, and assist in the development of extensions and new features.

The current 5G testbed consists of two next-generation base stations, three mobile equipments, a centralized unit, and a mobile core network. Work is in progress to virtualize and isolate the different software functions, as recommended in 5G and beyond.

The student will have the opportunity to work with a fully-functional 5G mobile network and help in the testing and developing of new features.

C/C++, 4G mobile networks (LTE), version control (Git), virtualization technologies. Optionally: Python.

Alberto Martínez Alba: alberto.martinez-alba@tum.de

Patrick Kalmbach: patrick.kalmbach@tum.de

5G networks are anticipated to support the tremendous growth in the traffic demands and the heterogeneity of future applications. In that regard, novel techniques such as Software-Defined Networking (SDN), Network Function Virtualization (NFV) pave the way towards programmable and flexible networks. The introduction of SDN/NFV advocates the notion of Network Slicing and Multi-access Edge Computing (MEC). On the one hand, network slicing is a promising approach for 5G to facilitate the deployment of multiple heterogeneous networks on the same physical infrastructure. On the other hand, MEC can be utilized to reduce the potential service delays. In this thesis the student shall focus on the algorithmic part of network slicing and on the performance advantages possibly brought in by the introduction of MEC. To this aim, machine learning approaches are expected be utilized for the cache replacement in order to boost the network performance.<

 This master’s thesis is a collaboration with German Aerospace Center (DLR). In this master thesis the student will  survey the most recent network architectures integrating satellite and terrestrial 5G elements in order to understand the role played by Network Slicing for end-to-end network resource allocation and assess the performance of an exemplary MEC solution in the proposed scenario. Moreover the focus will be on the investigation of the impact of time-varying popularity on caching strategies. 

•  Good knowledge of a simulation environment Matlab, Python.
•  Good mathematical background. Basic knowledge of satellite and cellular communications.

Optical core networks deploy fixed grid in order to accommodate their demands. The trend is however, migrate towards flexgrid networks able to use as much spectrum as required by the demand allowing a better use of the spectrum. However, the Routing and Wavelength Assignment (RWA) paradigm becomes more complex as it has to find the suitable spectrum of the demand.

This thesis will consider two different alternatives to solve the Routing and Spectrum Assignment (RSA) problem. Given a database providing for each (source, destination) pair a set of (route, modulation format, baud rate, OSNR, ….), choose the best route and spectrum for a set of demands.

The implementation will be done in Java. The first alternative is based on an existing work (JOCN2018 paper of Soumplis et al.) and the second alternative will be proposed during the thesis.

The evaluation aims at comparing both solutions in terms of computation time, network utilization, network balancing, blocking probability, energy, etc. Different scenarios will be studied: i) Static, ii) Dynamic sequential: ordered set of demands with Tstart, iii) Dynamic random: set of random demands with (Tstart, Tend)

Java programming skills

Basic knowledge of flexible optical networks

Wireless Reliability Certification is an unknown process that will be a biggest part of future automation networks. It is currently, unclear why certain standards cannot achieve certification when some do. 

In this work, we plan to do a state of the art search on reliability certification for wireless technologies also including wired certification for comparison.

We plan to incorporate a small simulation to show due to which characteristics a wireless technology may not comply with this certification process.

P4 is a novel programming language for programming networking devices (switches and NICs). It allows you to describe how the networking devices will process the packets in a high-level language. This programmability paved the way to many different applications and innovations that were not possible before, which made it a hot research topic nowadays. The objective of this Master Thesis/internship topic is to evaluate the performance of such devices. The main scope of this thesis/ internship work is to measure and model the performance of P4-programmable devices. Note that there are multiple related subtopics, so the required tasks are flexible and will be based on student's interest and skills. The main requirements are being motivated and having basic programming and networking skills. Knowledge about SDN and P4 is a plus. 

Motivation, Basic Programming and Networking skills

LiFi (visible light communication) is becoming an increasingly important technology as the unlicensed ISM band gets overcrowded. LiFi, with its high data rates, emerges as a promising solution to provide connectivity in an aircraft using overhead LED lights thereby reducing the interference with other wireless technologies and reducing the overall aircraft radio emissions. 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 WiFi cells. 

In this work, the student is expected to develop modules necessary for the simulation of a LiFi network in NS3 and use existing WLAN modules to simulate an in-cabin heterogeneous LiFi-WiFi network to analyze the attainable data rates, reliability and other QOS parameters. 

Related Reading:

1. H. Haas, L. Yin, Y. Wang and C. Chen, "What is LiFi?," in Journal of Lightwave Technology, vol. 34, no. 6, pp. 1533-1544, March15, 2016.
2. M. Ayyash et al., "Coexistence of WiFi and LiFi toward 5G: concepts, opportunities, and challenges," in IEEE Communications Magazine, vol. 54, no. 2, pp. 64-71, February 2016.

• Object-Oriented Programming C++, Python
• Experience with NS3 module development preferable
• Sound knowledge of Wireless Communication, especially WLAN

Sensor to Cloud Solution via a Gateway

You are wondering how future 5G networks will operate? Your are able to read? You want to know what the acronyms in the keywords mean? Then check out the publication "NFVnice: Dynamic Backpressure and Scheduling for NFV Service Chains".

• C
• C++
• Python
• Networking

Das Thema dieser Bachelorthesis ist die Entwicklung, Optimierung und Evaluation des proprietären Brandmeldebussystems LSNi1 der Firma Bosch Sicherheitssysteme GmbH in Grasbrunn.

Brandmeldeanlagen sind ein bestehender notwendiger Teil der Gebaudetechnik und sollen fur neue vernetzte IoT-Dienste verwendet werden. Aufgrund immer höheren Anforderungen dieser Dienste soll die Datenrate uber den Brandmeldebus gesteigert werden. Zur Zeit wird das Bussystem zur verhaltnismäßig datenarmen Alarmabfrage und Antwort der Netzelemente verwendet. Ein wichtiger Punkt ist dabei, die Systemcharakteristika des Brandmeldesystems nicht zu verschlechtern. Solche Netzwerke sind als Sicherheitssysteme deklariert und mussen daher äußerst zuverlässig in Sachen Ausfallsicherheit und Fehlererkennung sein.

Die Aufgabe besteht darin, eine neue Übertragungstechnik mit höherer Datenrate auf der physikalischen Schicht des Busses zu finden und diese mit Hilfe eines Prototypen umzusetzen. Die Optimierung des Prototypen soll durch Berechnung und Experimentieren mit Systemelementen ausgearbeitet werden. Daraufhin werden verschiedene Testaufbauten, im firmeneigenen Labor getestet. Die dabei aufgenommenen Daten werden über ein entwickeltes Matlab-Evaluierungsprogramm ausgewertet, um eindeutige Aussagen über die Funktionalitat des Systems zu geben. Eine besondere Herausforderung bei Systemen dieser Art sind die hohen Kabellängen und die große Anzahl an Systemelementen auf dem Bus. Mit den ausgewerteten Ergebnissen soll überpruft werden, in welchem Umfang die Datenrate auf dem Bus gesteigert werden kann.

With the move towards NFV, network functions such as firewalls, load-balancers, NAT etc. are increasingly deployed as software. On the other hand the capabilities of networking hardware is not fully utilized in this architecture.

In this work the popular NFV framework click shall be enhanced with offloading capabilities to NICs. More specifically the NICs' matching capabilities shall be utilized.

For the work the following steps shall be completed:

• Implementation of an offloading module in the Click Framework
• Evaluation using functional and performance tests

• Basic networking knowledge (TCP/IP,etc)
• Programming in C and C++

The radio frequency spectrum is a scarce resource. As a result, particular parts of the spectrum are used heavily whereas other parts are vastly underutilized. For example, the unlicensed bands are extremely over utilized and suffer from cross-technology interference. Monitoring and understanding the spectrum resource usage is critical in order to improve and regulate the radio spectrum utilization. Currently neural networks has been used with cognitive radio to identify modulation type, source of interference and interference free frequency channels from spectrum data.
The goal of this thesis is to characterize the unlicensed spectrum as much as possible to have a better understanding of the spectrum environment. For example some features could be to identify the busy channels, their duty cycles, nearby Bluetooth, microwave or other devices operating in same frequency band, narrowband interference, white spaces etc.
The first part of the task involves using a USRP along with GNU radio to study the characteristics of waterfall spectrum for the above condition. Then the characterization process could be automatize using some algorithm. The best choice of algorithm would be identified during the thesis.
A possible approach could be using Neural Networks. Like train a network using previously studied (Training) spectrum data so that it can characterize a new one with a certain accuracy. However NN may not always be the ideal solution for every case so other image recognition algorithm may be applied.

High reliablity and low latency requiring applications, namely Industrial IoT, (IIoT) cannot be supported by the current version of the mobile networks standard that is the 5G. 5G also known as New Radio, is currently on going through the standardization process and one of the critical topics is to support IIoT applications. 

There are many solutions proposed for radio resource management to incorporate IIoT in 5G. Modern Random Access techniques are the most prominent solution for efficient radio resource management through use of interference cancellation. 

In this work the main Modern Random Access algorithms are optimized for low number of users and high reliability and low latency constraints. 

Softwarized wireless sensor networks base station to prototype 5G-like ecosystem using cheap sensor to test out most recent radio resource management techniques.  

yle="margin: 0px 0px 1em; color: #333333; font-family: Arial, sans-serif; font-size: 14px;">Network slicing is a novel approach used in 5G networks to enable the deployment and coexistence of heterogeneous networks with different Quality of Service requirements in the same physical infrastructure. In the Radio Access Network (RAN) the network slicing problem is complex due to the stochastic nature of wireless channels and scarce resources. To this end efficient algorithms are required. This has attracted a lot of attention in the literature where there are works dealing with RAN slicing and QoS provisioning but focusing on a single cell scenario. However, for specific applications (i.e., aircraft in-cabin communication, reliability critical tasks) one cell might not be sufficient. Therefore, the need of multiple-cells becomes mandatory.

yle="margin: 0px 0px 1em; color: #333333; font-family: Arial, sans-serif; font-size: 14px;">In this master thesis the student will focus on the downlink scenario of a cellular system consisting of multiple-cells. The focus will be on the dynamic allocation of resources with regard to slicing requests for guaranteeing QoS and isolation among them. Moreover, an optimization problem has to be developed analytically and further evaluated with extensive simulations to prove the efficiency of the approach. Finally, for specific metrics such as reliability this method shall be compared with existing state-of-the-art approaches.

yle="list-style-type: square; margin-left: 0px; padding-left: 20px; color: #333333; font-family: Arial, sans-serif; font-size: 14px;">
• yle="padding-left: 0.3em;">Good knowledge of a simulation environment Matlab, Python
• yle="padding-left: 0.3em;">Good mathematical background. Knowledge of optimization techniques and tools such as Gurobi is a plus.

TBD

In this thesis student has to implement a switch module for enabling byzantine fault tolerance (BFT) in SDN using P4.

Current mobile devices support several wireless technologies to communicate with the outer world. To this end, these devices often include multiple LTE, WiFi, or BlueTooth interfaces. In addition, base stations also have several wired or wireless interfaces to communicate with one another or with the core network. This adds a new degree of freedom to the problem of optimally selecting data flows within a mobile network, which is motivated by the ever-changing nature of wireless channels. In this thesis, a flexible mobile network supporting live selection of interfaces will be analyzed. Available algorithms to adaptively select interfaces and reconfigure data flows will be explored. A flexibility framework will be used in order to highlight the importance of request changes and reaction time.

Familiarity with mobile and wireless networks. Experience with MATLAB, Python or other programming languages or simulation environments. 

Multi-path routing is an important enabler of load balancing, resilience and traffic engineering in communication networks. However, in practice carrier networks are rarely deploy these methods owing to the required architectural changes, such as the application of shim headers, deployment of middleboxes, and hence, the lack of backward compatibility. The task of the student is to develop and implement a multi-path routing architecture by extending the current OSPF routing protocol with the ability to choose between multiple (two) next hops based on the value(s) of the traditional IP header fields and the available OSPF routing information.

Testbed with QT GUI and 4 mini USRPs for communication. GNURadio implementation for the SDRs.

Motivation für das Thema:

Hier ist die Idee, das wir mittels eines adaptiven, auf Reinforcement Learning basierenden, Routing Prinzip (Stichwort Q-Routing) das Routing in sich schnell ändernden ad-hoc Netzwerken bei der Car2Car Kommunikation im Vergleich zu bisher existierenden Routingverfahren verbessern wollen. Zu diesem Thema gibt es bereits Nachforschungen, die zu dem Schluss kommen, dass Q-Routing sich für diese Zwecke gut eignen sollte.

Problemstellung:

Bei dieser Arbeit ist das zentrale Problem, verschiedene, auf Reinforcement Learning basierende, Routingverfahren zu implementieren, gegenbenenfalls weiterzuentwickeln, und mit den gegenwärtig üblichen Routingverfahren im Geonetworking-Bereich (Stichwort Greedy-Routing, Congestion Based Forwarding) zu vergleichen.

Technische Umsetzung:

Zur technischen Umsetzung habe ich dir das Paper, was für uns als Inspiration dient, an die Mail angehängt. Neben Q-Routing überlegen wir auch darüber nach, auf SDN zurückzugreifen, beispielsweise in Form eines Controllers, der auf einem Mobile Edge Cloud Knoten läuft. 

Wissenschaftlicher Beitrag:

Der Hauptbeitrag von diesem Thema wäre vor allem die Untersuchung dieses Routingverfahrens in einer realistischen Umgebung. 

The goal of this thesis is the implementation and evaluation of a deep reinforcement learning based approach to traffic engineering in datacenters. Your task is the implementation of a packet-level simulator in NS3 and the integration of our in-house deep-rl framework in this simulator. Further, you should evaluate different neural architectures as well as data-center topologies and compare the performance obtained with deep RL to Equal Cost Multi-Pathing (ECMP).

High reliablity and low latency requiring applications, namely Industrial IoT, (IIoT) cannot be supported by the current version of the mobile networks standard that is the 5G. 5G also known as New Radio, is currently on going through the standardization process and one of the critical topics is to support IIoT applications. 

There are many solutions proposed for radio resource management to incorporate IIoT in 5G. Modern Random Access techniques are the most prominent solution for efficient radio resource management through use of interference cancellation. 

In this work the main Modern Random Access algorithm, Irregular Repetition Slotted ALOHA (IRSA), is adapted for 5G New Radio through new flexible numerology. On top of this a multichannel adaptation to support low latency with IRSA is proposed.

Measurement of the control and management traffic of within a OpenStack deployment.

Enabling virtualization on top of programmable data plane described by P4 language.

Within an SDN network, in order to avoid filling up the flow tables of switches, routing algorithms have to keep track of the amount of rules installed at every switch. When installing a new flow, such algorithms try to minimize the amount of additional rules required. To do so, a solution is to run a flow rules aggregation algorithm for each switch visited by the algorithm. As a result, the routing algorithm is aware of the amount of rules required to embed the new flow and can accordingly optimize its routing decisions.

The goal of the thesis consists in the implementation, evaluation and proposal of flow rules aggregation algorithms to be used within delay-constrained routing algorithms. In such a context, as the algorithm is run each time the routing algorithm visits a node, runtime is a very important dimension.

P4 is a novel programming language for programming networking devices. It enables the network engineer to describe the behavior in a simple language. Switches via P4 still needs to be communicated with a controller as in any SDN architecture, however, currently available. The latter is the reason for designing a new southbound interface, "P4 Runtime". The objective of this project is to evaluate the performance of the P4Runtime protocol and extend the existing SDN controller's benchmarking tools. A detailed analysis of the results will be performed. The student wants to work on this project in collaboration with Nokia Bell Labs and get a great experience on the trends in networking and SDN. In addition, 

Motivation and good analysis skills, Programming skills: JAVA C ++, SDN and programmable data plan knowledge is a plus

You are interested in the internal workings of a data-center? You want to get a head-start in programmable data-planes? You want to work on an actual problem? Then this is your chance!

Load balancing and traffic engineering within data-centers is still an open problem (belief it or not). Programmable data-planes offer new possibilities in tackling those problems. However, the topic is new and not well understood yet. Check out the publication "HULA: Scalable load balancing using programmable data planes" to get an idea of what is awaiting you.

Your job in this project is the re-implementation and evaluation of an in-data-plane load-balancing algorithm. You will obtain a deep understanding of the fundamental problems in data-center networking and novel technologies such as P4 and SDN.

You also want to do a Forschungspraxis or Master Thesis? No problem! This project allows you to continue by bringing machine learning into play. Our goal is a network that learns traffic engineering by itself without human intervention. Just let us now.

• C++
• Discrete Event Simulation
• Python

In this work, the student will survey different P4 software switches and evaluate its performance.

This work studies the application of Boltzmann machines as an extension of Hopfield neural networks for optimization problem space reduction. This work should survey, analyze, and implement a first approach of a Boltzmann machine for an optimization problem. Potential extensions can also be applied to ongoing research studies of reconfigurable networks or virtual network embedding.

Python, Neural Networks, Mathematics

The student is provided with a neural architecture with the purpose to learn artificial features for a network that faciliate navigation in the network.

The student should train different architectures and evaluate them to verify or falsify the hypotheses that such features can be learned and how valuable the learned features are.

- Tensorflow

- Python

A routing configuration that guarantees high performance for specific traffic conditions can fail miserably in maintaining good performance when traffic conditions change. Traditional routing schemes address this challenge in two manners: 1) Optimizing routing with respect to past traffic conditions, with the hope, that good routing configurations for past-observed traffic conditions will also fare well in the future 2) optimizing static routing configurations with respect to a wide range of feasible traffic scenarios. Machine Learning based approaches present a third option: leveraging (implicit) information in past traffic conditions to learn good routing configurations for future conditions. But how can a learning model acquire the knowledge to 1) determine a good routing configuration to begin with and 2) leverage the past knowledge? Reinforcement Learning is a viable solution in this context. The agent learns through interaction with the network what good routing configurations are and how to leverage past information to do so. The goal of this thesis is the re-implementation of "Softmin Routing" presented in [1]. In this approach, a series of past traffic matrices is used to obtain edge weights, which are then converted to splitting ratios in destination based forwarding. Softmin Routing should then be evaluated on different data-center topologies and for different workloads. [1] A. Valadarsky, M. Schapira, D. Shahaf, and A. Tamar, “A Machine Learning Approach to Routing,” CoRR, vol. Abs/1708.03074, 2017.

• Machine Learning
• Traffic Engineering
• Python

Survey of software switches. Setup of physical and virtual machine-based measurement system. Code analysis of Open vSwitch user space. Reevaluation of existing Open vSwitch measurements. Study of Bayesian optimization concept.

Python, C++, Linux

Network Function Virtualization (NFV) is becoming a promissing technology in modern networks. A challenging problem is determining the placement of Virtual Network Functions (VNFs). In this work, we plan to implement existing algorithms for embedding VNFs chains in NFV-enabled networks. 

Experience in Python or Java, object oriented programming

amir.varasteh@tum.de

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 every Tuesday from 15:00 to 18:00.

Solid knowledge in Wireless Communication: PHY, MAC, and network layers. Solid programming skills: C/C++. Linux knowledge. Experience with embedded systems and microcontroller programming knowledge is preferable.

The wireless sensor networks lab offers the opportunity to develop software solutions for wireless sensor networking systems, 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 every Tuesday from 15:00 till 18:00.

The yle="color: #333333; font-family: Arial, sans-serif; font-size: 14px;">Software-Defined Networking (SDN) lab offers the opportunity to work with real hardware on interesting and entertaining projects related to the SDN network paradigm. For the next semester, a position is available to assist the teaching assistants for the lab (definition and preparation of the assignments, preparation of the hardware, etc.).

yle="list-style-type: square; margin-left: 0px; padding-left: 20px; color: #333333; font-family: Arial, sans-serif; font-size: 14px;">
• yle="padding-left: 0.3em;">Solid knowledge in computer networking (TCP/IP, SDN)
• yle="padding-left: 0.3em;">Solid knowledge of networking tools and Linux (iperf, ssh, etc)
• yle="padding-left: 0.3em;">Good programming skills: C/C++, Python

This working student position supports us in conducting measurement and implementing network algorithms and solutions for various use cases.

Machine Learning (ML) and Artificial Intelligence (AI) are also a current hype in networking research. Within, the applications of ML and AI range from data center network optimization to anomaly detection in campus and enterprise networks. As we are currently conducting a lot of research in this area of application of ML/AI, we are always looking for students willing and able to support us. So, if you are interested in ML/AI and its application to networking, please contact us.

We offer you:

• Possibility to extend your skills in ML/AI, programming and networking

• Possibility to become (Co)-author of publications already before graduating

• Access to compute cluster and data center-like networking environments

We need:

• A student with good programming skills in Python and C++

• Basic knowledge in data science

• Basic understanding of networking: TCP/IP, UDP, BGP, RIP, FatTree Topologies, ...

• Basic knowledge in algorithms and optimization: search, sort, mathematical problem formulation, …

• Basic skills in Linux and server administration

Python, C++, Data Science, Basic Networking