Contact
Director
Prof. Gerhard Fohler
Secretary
Stephanie Jung
Phone: +49 631 205 2086

Office Hours
Mon/Tue: 13:00-16:00
Wed-Fri: 08:00-11:00
You are here: RTS-Portal>>Publications

Publications

Search Publications

show/hide

Dimensioning buffers for AFDX networks with multiple priorities virtual links


Category: Inproceedings
Author(s)Rodrigo Coelho, Gerhard Fohler, Jean-Luc Scharbarg
SourceDigital Avionics Systems Conference, IEEE Computer Society, Prague.
Date Sep 2015
Download:

Abstract

AFDX allows for bandwidth reservation guaran- tees by means of virtual links and permits the classi- fication of virtual links in one of two priority levels: high or low. AFDX compliant switches offer output buffers at each switch output port and the stored frames leave each output port according to a fixed priority FIFO policy. Overflow of these buffers must be avoided at all cost to prevent data loss. Although the AFDX standard determines the minimum buffer size dedicated to an output port (we will refer to the output port as a node), the actual length of each priority buffer (high and low), is a designer decision. Previous work has presented methods to compute upper bounds for the buffer backlog of AFDX nodes where all virtual links have the same priority level. This paper presents a method for dimensioning the buffers of AFDX nodes considering multiple priorities virtual links. We compute an upper bound for the buffer backlog on each node and assume that each virtual link may be classified with any priority out of a set of priority levels: two priority levels, as specified on the AFDX standard, or more. In this paper we assume that all frames on the network have the same size, e.g., their size is equal to the size of the largest frame. Our method considers the worst case busy period of each priority level of each node (computed for instance by means of the trajectory approach) and extracts the number of frames of each virtual link that impact on the buffer backlog of that priority level. We then identify the arrival order of frames that leads to the largest backlog of the priority under analysis and compute the upper bound.

Bibtex


Upcoming Events
Currently no events
Latest News:
04.09.17 11:03
Our paper titled "Modeling Offset Adaptations for Latency Minimization in Cascaded...


03.07.17 10:13
Our paper titled "DREAMS Toolchain: Model-Driven Engineering of Mixed-Criticality...


17.06.17 12:08
20.06.2017, 11:45-13:15 @ 11-207 on "Microkernel-Based Systems" by Dr. Marcus Völp (SnT –...


All News