Dataset for loosely coupled hybrid scheduling of processing and communication in TSN-based IMA systems

Citation Author(s):: Xuan Zhou
Submitted by:: Xuan Zhou
Last updated:: Tue, 12/26/2023 - 02:28
DOI:: 10.21227/xcxb-n346

138 views

Categories:

Communications

Keywords:

Integrated modular avionics

Time-Sensitive Networking

partition

time-triggered schedule

end-to-end delay.

ACCESS DATASET CITE

Abstract

This dataset serves loosely coupled hybrid scheduling of processing and communication for TSN-based IMA systems. It includes 8 experimental case configurations with 3 structures (mesh, ring and tree) and 3 scales (small, medium and large), and the cases are named with corresponding structure and scale, such as Mesh-structure Small-scale (MS) case, Mesh-structure Medium-scale (MM) case, Mesh-structure Large-scale (ML) case, Ring-structure Small-scale (RS) case, Ring-structure Medium-scale (RM) case, Ring-structure Large-scale (RL) case, Tree-structure Small-scale (TS) case, Tree-structure Medium-scale (TM) case, and Tree-structure Large-scale (TL) case, where MS case and RS case are the same. It also includes the detailed experimental results for scheduling designs based on these cases.

Instructions:

Basic data structure:

In this dataset, the data is saved in the form of JSON files for easy reading and writing. For the TSN-based IMA systems, object classes mainly include node, link, partition, task and messages. Their related keys are as follows.

Node: identifier, category (switch or module), con_partition (the identifier list of all partitions in this node), hyper_period (the hyper period of all partitions in this node), utilization (the proportion of partition activation time in this node).
Link: identifier, src_node (the identifier of its source node), de_node (the identifier of its destination node), con_message (the identifier list of all messages traversing this link), hyper_period (the hyper of all messages traversing this link), utilization (bandwidth utilization of all messages).
Partition: identifier, module (the identifier of the module in which this partition resides), period, length, con_task (the identifier list of all tasks in this partition), state (the notation of whether this partition has been scheduled, ‘done’ means this partition has been scheduled, and ‘undone’ means this partition cannot be scheduled), offset (the scheduling result, default value is -1).
Task: identifier, partition (the identifier of the partition in which this task resides), release (the release time of this task, default value is 0, support further model improvement), period, length, deadline, priority (default value is -1, dynamically assign value during the scheduling, support further model improvement), state (the notation of whether the bounds of this task has been calculated), bound (‘wcrt’ is the worst-case response time, ‘bcrt’ is the best-case response time, ‘wcst’ is the worst-case release time, and ‘bcst’ is the best-case release time, note that for task-message hybrid scheduling methods, ‘bcst’ and ‘wcst’ are both the scheduling offset of this task, ‘bcrt’ and ‘wcrt’ are both the sum of the scheduling offset and the length of this task).
Message: identifier, src_node (the identifier of its source node), des_node (the identifier of its destiniation node), src_partition (the identifier of its source partition), des_partition (the identifier of its destination partition), src_task (the identifier of its producer task), des_task (the identifier of its consumer task), period, length, mode (queue or sample, support further model improvement), deadline, bandwidth, state (the notation of whether this message has been scheduled, ‘done’ means this message has been scheduled, and ‘undone’ means this message cannot be scheduled), inject (the injection time of this messages, also the scheduling result), path (the identifier list of the links that this message traverses), offset (the offset of this message on corresponding links, also the scheduling result), delay (‘network’ is the network-layer ETE delay of this message, ‘application’ is the application-layer ETE delay of this message).

Case_configuration Folder: the original configuration of all cases, where node.json only includes identifier and category, link.json only includes identifier, src_node and des_node, partition.json only includes identifier, module and length, task.json only include identifier, partition, period and length, message.json onlye includes identifier, src_task, des_task and length.

Case_initialization Folder: the initialization configuration of all cases after automatically completing all required information for scheduling design.

Comparison_algorithm Folder: the scheduling results of different algorithms, where PMNP and PMWP are our proposed partition-(task)-message scheduling algorithms, MQO is an optimized message scheduling algorithm, MES is a (task)-message scheduling algorithm, TMPR and TMSC are two task-message scheduling algorithms.

Comparison_parameter Folder: the scheduling results of our partition-(task)-message scheduling algorithm with different parameters, including injection mode (centralized and distribute), transmission rate (100Mbps, 125Mbps, 200Mbps, 250Mbps, 300Mbps, 400Mbps, 500Mbps, 1000Mbps), slot length of ICT (0.01ms, 0.1ms, 0.2ms, 0.5ms, 1ms, 2ms), and slot length of GCL (1ns, 5ns, 10ns, 50ns, 100ns, 200ns, 250ns, 500ns, 1000ns, 2000ns, 2500ns, 5000ns, 10000ns, 20000ns)