This page is not available in your selected language. Falling back to default.

Implementation of local deadline assignment for distributed systems with shared resources

Project type: Hiwi-Job, Bachelor Thesis, Master Thesis
Status: open

In a distributed real-time system (DRTS), applications are decomposed into tasks, which are then assigned to processors according to locality constraints. Often because the system designers are forced to modularize software for maintainability and reusability reasons or exploit parallelism for performance reasons, tasks have usually been logically decomposed into a set of sequential and/or parallel subtasks, each one having processing demands at some node in order. To guarantee the functionality of a real-time system, a distributed application task often requires end-to-end guarantees on quality-of-service (QoS). For example, in hard distributed real-time systems, an application task is subject to timing constraints typically specified by end-to-end deadlines, and the optimization goal of such a distributed system is often defined as maximizing schedulability (i.e., maximizing the number of schedulable tasks). 

In this work, the student will explore the local deadline assignment for DRTS with shared resources. The practical part of the work is to implement different local deadline assignment algorithms for DRTS with shared resources. The results will then be compared with the ones obtained with our newly developed methodology.

 Knowledge requirements:

  • Programming knowledge (e.g. C,  C++, JAVA, Python)
  • Analytical skills
Hans-Sommer-Street 66
38106 Braunschweig

Fon: +49 (0)531 391-3734
Fax: +49 (0)531 391-4587

Address of Prof. Jukan's workgroup:

Hans-Sommer-Street 66
38106 Braunschweig

Fon: +49 (0)531 391-5286
Fax: +49 (0)531 391-5296