Autonomous Control and Management in Heterogeneous Networks
news
06/ 09/ 2010
miniCMS home
is now on this page.
06/ 08/ 2010
Time Lapse Video
from the Lange Nacht der Wissenschaften online!
05/ 26/ 2010
Meyyar and Dipak
join the team as student assistants. Welcome!
Open Thesis/ Student Assistant positions
Autonomously Migrating Services
We are currently working on a platform that brings autonomous functionality into the home context. The autonomic capabilities are realized in form of control services that run on nodes with sufficient resources inside the network. As the services can provide (system) critical functionality, like the control of building security equipment, it has to be ensured that they will not fail when nodes fail. It is also desired that the services run in an optimal way, so on nodes that currently have sufficient free resources.
Our functional services are logically situated on top of our already existing knowledge-based middleware. As it already makes location of information and devices transparent to the services they can be moved freely in principle.
In this work functionality to run and move services dynamically inside the platform should be specified and implemented.
The following questions should be answered in this thesis:
Which approaches of moving services inside a network do exist and are currently in use?
Which are existing algorithms for deciding on which node to place a service?
Which are existing algorithms for splitting and merging services?
Which are the domains this functionality is used today?
What are the prerequisites of the respective approaches?
Concerning the specification:
Which components are needed?
How can the solution be realized in the most decentralized way?
Which base functionality is needed for a service to be movable, splittable and mergable?
How can it be ensured that a service moves with no or minimized interruption?
How can the overall load be balanced?
The specified functionality should be implemented on top of the knowledge middleware. The result should be evaluated based on test scenarios and measurements.
Our functional services are logically situated on top of our already existing knowledge-based middleware. As it already makes location of information and devices transparent to the services they can be moved freely in principle.
In this work functionality to run and move services dynamically inside the platform should be specified and implemented.
The following questions should be answered in this thesis:
Which approaches of moving services inside a network do exist and are currently in use?
Which are existing algorithms for deciding on which node to place a service?
Which are existing algorithms for splitting and merging services?
Which are the domains this functionality is used today?
What are the prerequisites of the respective approaches?
Concerning the specification:
Which components are needed?
How can the solution be realized in the most decentralized way?
Which base functionality is needed for a service to be movable, splittable and mergable?
How can it be ensured that a service moves with no or minimized interruption?
How can the overall load be balanced?
The specified functionality should be implemented on top of the knowledge middleware. The result should be evaluated based on test scenarios and measurements.
