Schedule and Overhead Trace of Mixed Critical Applications using Litmus-RT

Citation Author(s):
Submitted by:
Preeti Godabole
Last updated:
Fri, 04/12/2024 - 12:57
Data Format:
0 ratings - Please login to submit your rating.


<p>Mixed critical applications are real-time applications that have a combination of both high and low-critical tasks. Each task set has primary tasks and two backups of high-critical tasks. In the work carried out, different synthetic workloads and case studies are used for extracting the schedule and overhead data on a real-time operating system. The utilization of the task sets lies between 0.7 to 2.4.The Linux kernel is recompiled with Litmus-RT API to monitor the scheduling decisions and also measure the overheads incurred. Litmus-RT API is used to monitor the schedule decisions and overheads incurred during scheduling on Linux-based machines. The schedule decisions are monitored on a quad-core system for partitioned-EDF scheduling (P-EDF). Two processor variants namely ARM processor and Intel processor are used to extract the data set.&nbsp; The ARM cortex processor with 600MHz speed on R-Pi model B is used in the research work. Another variant used is the Intel i7 processor with a speed of 2.8GHz. Both the systems used have homogeneous quad-core processors.&nbsp;</p><p>The task sets are run on the operating system and the scheduling algorithm used is P-EDF with three different allocation algorithms proposed by the researchers to achieve fault tolerance. The data set consists of task sets used along with allocation tables, schedule trace, and overhead trace files for all the task sets taken for experimentation.</p><p>&nbsp;</p>


The dataset-schedule-overhead has two zip files, rpi and intel denoting the processor on which the data is collected. The rpi zip file has 19 Excel workbooks and intel file has 18 excel workbooks. Task set workbook has the details of all the mixed critical task sets used for experimentation. The names of the workbook beginning with rpi indicate that the trace files are taken on Raspberry-pi and the names beginning with Intel indicate the data is extracted on an Intel processor. Every workbook has multiple sheets. The sheet named task-set gives the pid and the task mapping. The worksheets with names starting with schedule indicate they are schedule trace files and the ones starting with overhead indicate they are overhead trace for that respective task set. The trace files are segregated on the basis of processors utilized for executing the task sets.
1. Nomenclature used to name the workbooks – rpi-instru-clust, first part gives the processor variant, second part gives the name of the task set and third part is the name of the allocation algorithm.
2. schedule-intel-p0 worksheet in the workbook intel-instru-clust gives the scheduling decisions on processor P0 of type Intel.
3. Overhead-intel-p0 worksheet in the workbook has the overheads measured during the scheduling of the task sets.
The schedule trace file has 11 attributes and overhead trace has 8 attributes. The details and meaning of the attributes can be obtained on the link:

Funding Agency: 
Higher and Technical Education Department, Government of Maharashtra, Maharashtra State, India
Grant Number: 
TVY 1423/ 83/23/TE-3 Dated July 07, 2023


This data set is extracted for mixed critical systems on real time operating systems.

Submitted by Preeti Godabole on Fri, 04/12/2024 - 12:59