Skip to content

xyjiang1216/rtss_paper_131

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Introduction

The source code of RAP-Plan, Window and Frame algorithms used by RAP: A Policing Mechanism Guaranteeing Sub-microsecond Determinism against Abnormal Traffic in TSN in RTSS 2023 paper #131.

Getting Started

  • Package installation
pip install -r requirement.txt

Run

  • Run the file in the \bin directory to run the corresponding algorithm, e.g.,
python main_for_RAP_demo.py

Constraints in Phase II of RAP-Plan

1. Period Constraint.

This constraint requires that the TS frame of a period must be scheduled during the current period. There are two reasons. First, this constraint could reduce the buffered frames in the switch. If the frame of (k+1)-th period has arrived while the frame of the k-th period has not been scheduled, then the switch needs to buffer two frames. Yet the on-chip memory is limited. Second, the constraint can reduce the search space to a reasonable range. This constraint corresponds to the Frame constraint in citation#18.

Period Constraint

2. Sequence Constraint.

The routing of a TS frame is sequential. The scheduling time on the upstream link must be earlier than the scheduling time on the downstream link. Besides, the scheduling time between neighbor links must be larger than the worst processing and transmission delay. Otherwise, the frame has not arrived at the downstream link but the link is already scheduling the frame. The rst has been lengthened in Phase I, thus the scheduling time of downstream only needs to be one rst larger than the scheduling time of the upstream link. This constraint corresponds to the Flow Transmission constraint in citation#18.

Sequence Constraint

3. Queue Resource Constraint.

The number of queues in the switch is typically less than 8. The queues that can be used by TS frames are limited. This constraint corresponds to the last sentence Section-4-2 in citation#18.

Queue Resource Constraint

4. Deadline Constraint.

TS streams have a strict end-to-end deadline. The planned end-to-end delay must be less than the allowed maximum end-to-end delay. This constraint corresponds to the End-to-End constraint in citation#18.

Deadline Constraint

5. Contention-free constraint.

Contention-free constraint ensures that frames through the same link cannot use the same raster.

Contention-free constraint

6. Zero-aggregation constraint.

If frames enter the same queue at link (a,b), once the gate of this queue opens at some raster, frames in this queue will be scheduled at the same raster according to TAS. This obeys the first sub-goal above. Therefore, there should be only a single frame in a queue at the same time, i.e., zero-aggregation.

To avoid aggregation, the scheduling time for any two TS frames from upstream links can only be one of the following cases: (1) After the former frame has been scheduled to the downstream link, the latter frame is just scheduled at the upstream link; (2) The two frames use different queues. The formulation of the first case is as Eq.(6).

Zero-aggregation constraint

We represent Eq.(6) as $\Pi_{i,j}^{(a,b)}$. The complete constraint is shown in Eq.(7).

Zero-aggregation constraint

Contention-free constraint along with zero-aggregation constraint collaboratively achieves the first sub-goal.

7. Single-frame-per-raster constraint.

To achieve the second sub-goal, this constraint guarantees that frames from different upstream links $(x,a)$ and $(y,a)$ arrive at the current link $(a,b)$ at different rasters.

Single-frame-per-raster constraint

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages