Resource Sharing Under a Server-based Semi-Partitioned Scheduling Approach
Ref: CISTER-TR-141008 Publication Date: 8 to 10, Oct, 2014
Resource Sharing Under a Server-based Semi-Partitioned Scheduling Approach
Ref: CISTER-TR-141008 Publication Date: 8 to 10, Oct, 2014Abstract:
The rapid evolution of commercial multicore platforms has
raised the industry interest in developing and running applications
independently on the same platform. However, in
realistic industrial settings, tasks belonging to different applications
share resources that are not limited to the CPUs.
Applications share hardware components (e.g. co-processors
and actuators), and need to access portions of code that may
be protected by semaphores or mutexes.
In this paper we address the important challenge of resource
sharing on multicore platforms through the use of
servers, i.e, through a hierarchical scheduling approach, which
is an effective technique to ensure the integration of independently
developed applications on the same computing platform
as well as the isolation of tasks. To solve that problem
we adapt and extend the MrsP [6] resource sharing protocol
and further combine it with the NPS-F scheduling algorithm
[5], which employs a server-base approach. A schedulability
analysis is then provided for the resulting framework.
Alexandre Esper
Eduardo TovarDocument:
View
Download22nd International Conference on Real-Time Networks and Systems (RTNS 2014).
Versailles, France.
Record Date: 8, Oct, 2014