Decision Sciences & Systems
Technical University of Munich


 mg 4297

Contact Information

Email:  This email address is being protected from spambots. You need JavaScript enabled to view it.
Phone:  +49 (0) 89 289 - 17528
Fax:  +49 (0) 89 289 - 17535
Office: Room 01.10.058
Boltzmannstr 3
85748 Munich, Germany

Theses Topics

Please see also a complete list of other theses topics offered by our chair.

  • Topics
    • Topics are in the area of cloud computing and cloud platforms
    • All topics require good programming skills in Java and Python
    • Skills in C/C++ and Golang are beneficial
  • OpenStack Resource Allocation
      OpenStack has become one of the most important management tools for public (Rackspace) and private clouds. Still, it lacks control features like VMware’s Distributed Resource Scheduler (DRS) and Distributed Power Management (DPM) in order to reallocate virtual machines so that the number of active servers in a data center is minimized. At our chair we conducted extensive research on how to assign physical resources to VMs while minimizing active servers and meeting service level constraints. In this thesis this approaches shall be implemented to the OpenStack system so that the approaches can be applied in real-world data centers.
  • Efficient load generator for benchmarking cloud applications
    • For conducting experiments in a cloud testbed we require a workload generator that can simulate tens of thousands of users in a memory efficient way. Various Java based solutions like Rain or Faban exist. They are based on threads to simulate each user leading to a huge memory consumption due to thread overheads. An alternative are lightweight threads as they are used in Golang, stackless Python or Java Kilim. The goal of this thesis is to implement a workload generator prototype in Golang and Java Kilim to compare its performance to existing solutions.
  • Handling VM allocation transitions
    • In enterprise cloud data centers VMs are assigned to servers at the time they are created. The allocation of a server cluster can be recalculated in regular intervals to match the VM workload profiles and to reduce the number of required servers. A new allocation can be implemented by live-migrating VMs. However, it is not possible to live-migrate all VMs at once as this would cause massive resource-oversubscriptions. Hence, the live-migration steps to transfer one allocation into the next allocation have to be planned. This thesis should evaluate different approach to calculate such live-migration schedules.

Research Interests

  • Cloud Computing / Automated Data Center Management
  • Web Technologies
  • Software Engineering
Decision Sciences & Systems (DSS), Department of Informatics (I18), Technische Universität München, Boltzmannstr. 3, 85748 Garching, Germany
©2002-2020 DSS All Rights Reserved
Impressum, Privacy Policy, Copyright Information and Disclaimer
Cookies make it easier for us to provide you with our services. With the usage of our services you permit us to use cookies.