A Novel Approach to Achieving End-to-End QoS for Avionic Applications Electronic Communications of the EASST Volume 080 (2021) Conference on Networked Systems 2021 (NetSys 2021) A Novel Approach to Achieving End-to-End QoS for Avionic Applications Yevhenii Shudrenko, Daniel Plöger, Koojana Kuladinithi and Andreas Timm-Giel 5 pages Guest Editors: Andreas Blenk, Mathias Fischer, Stefan Fischer, Horst Hellbrueck, Oliver Hohlfeld, Andreas Kassler, Koojana Kuladinithi, Winfried Lamersdorf, Olaf Landsiedel, Andreas Timm-Giel, Alexey Vinel ECEASST Home Page: http://www.easst.org/eceasst/ ISSN 1863-2122 http://www.easst.org/eceasst/ ECEASST A Novel Approach to Achieving End-to-End QoS for Avionic Applications Yevhenii Shudrenko1, Daniel Plöger2, Koojana Kuladinithi3 and Andreas Timm-Giel4 1yevhenii.shudrenko@tuhh.de, 2daniel.ploeger@tuhh.de, 3koojana.kuladinithi@tuhh.de, 4timm-giel@tuhh.de Institute of Communication Networks, Hamburg University of Technology, Germany Abstract: Future Internet of Things (IoT) applications, such as connected industry 4.0, become more challenging with the strict Quality of Service (QoS) requirements, including reliability and delay guarantees. Several mechanisms in the communica- tion stack to match the expected QoS are already discussed and specified at different layers, with the goal to make the communication more reliable. They focus on the layer-specific enhancements. For example, Time-Slotted Channel Hopping (TSCH) is a link layer mechanism to avoid narrowband interference. On the network layer, several multi-path routing schemes are proposed to distribute the traffic load or to have backup paths with the purpose of making data transmissions more robust to link failures. In addition to the layer-specific improvements, an integration of the cross-layer in- formation can guarantee an end-to-end QoS for communication in dynamic envi- ronments. In this work we propose and evaluate a cross-layer framework for cell- disjoint routing, which eliminates overlapping resource scheduling in both time and frequency. It enables the end-to-end QoS for wireless sensor networks under the IPv6 Over the TSCH Mode of IEEE 802.15.4 (6TiSCH). The proposed framework, called 6TiSCH stack with cross-layer information exchange (6TiSCH-CLX), is val- idated on a selected set of aviation industry applications using both simulations and analytical model. Keywords: End-to-end QoS, WAIC, TSCH, RPL, Cross-layer, Avionic, 6TiSCH, Wireless sensor networks, Delay minimization 1 Introduction Modern aircraft uses different communication technologies and a large number of sensors and actuators for the purpose of improving the operations and simplifying the maintenance proce- dure. However, the current means of communication used for maintenance purposes, for example monitoring the temperature inside the cabin, are still based on wired connectivity. In 2015, the frequency band of 4200 to 4400 MHz was allocated for on-board wireless communication. The currently discussed standard is called Wireless Avionic Intra Communication (WAIC). Three typical scenarios which may utilize the WAIC standard are shown in Table 1. They have hetero- geneous requirements w.r.t. number of nodes, delay, traffic prioritization and prevalence. 1 / 5 Volume 080 (2021) mailto:yevhenii.shudrenko@tuhh.de mailto:daniel.ploeger@tuhh.de mailto:koojana.kuladinithi@tuhh.de mailto:timm-giel@tuhh.de A Novel Approach to Achieving End-to-End QoS for Avionic Applications Table 1: QoS requirements for WAIC simulation scenarios Name Latency (upper bound) Nodes Priority Periodicity Seat status 1s 200 Average aperiodic Smoke alarm 1s 30 High aperiodic Cabin humidity monitoring 60s 20 Low periodic WAIC requires strict transmission power limits to satisfy the highly demanding radio pro- tection criteria from [itu14]. While existing networks fail to meet the QoS requirements from Table 1, the approach presented in this work allows to adapt both routing paths and the link layer scheduling to support QoS demands of different applications. The proposed framework employs TSCH [tsc12], a Time Division Multiple Access (TDMA) channel access scheme in combination with the channel hopping, in the context of IPv6 Over the TSCH Mode of IEEE 802.15.4 (6TiSCH) stack for wireless sensor networks. The Routing Protocol for Low power and Lossy Networks (RPL) protocol enables the routing. Its capability to adapt to a tree based hierarchy reflects the topology in WAIC scenarios. The contribution of this work is twofold: Firstly, the proposed framework, called 6TiSCH stack with cross-layer information exchange (6TiSCH-CLX), is implemented and evaluated in the event-based OMNeT++ simulator with a full 6TiSCH stack and the Minimal Scheduling Function (MSF). The proposal provides RPL route selection and scheduling at the link layer, ensuring end-to-end QoS-aware communication by exchanging information across the protocol stack layers, in particular between the network layer, network sublayer and the link layer. Secondly, a mathematical model is developed to analyze metrics such as the end-to-end delay and Packet Reception Probability (PRP). The model allows to quickly obtain expectation and boundary values by varying parameters such as number of hops and link error rate. 2 Cross-Layer Framework The TSCH Medium Access Control (MAC) is designed to enable reliable communication in a scheduled manner. One slotframe in TSCH has a variable length and repeats periodically. Within a slotframe, one or more resources called cells, unique in both time and frequency do- main, are assigned to each pair of nodes. The TSCH standard does not specify the schedule management, so an appropriate scheduling is required for good network performance. On the contrary, improper scheduling can degrade the network performance significantly, causing cell collisions and delays. To address these issues, we propose a novel cross-layer framework, lever- aging RPL topology knowledge and a 6TiSCH Operation Sublayer (6top) sublayer to construct a collision-free TSCH schedule, while also targeting the QoS in terms of end-to-end delay. Despite numerous proposed centralized [PAD+12] and decentralized [DALW15] scheduling approaches, most of them focus on layer-specific improvements, rather than utilizing a hybrid, cross-layer strategy allowing for more flexibility and better control over the QoS. Collision Handling: The proposed framework operates in two main phases, aiming to elim- inate collisions and minimize end-to-end delay, respectively. In the first phase, cell coordinates NetSys 2021 2 / 5 ECEASST for nodes are calculated using their MAC addresses, same as in the MSF. Further, the RPL sink assigns each branch a unique subset of frequency channel offsets for enabling collision-free com- munication, as highlighted in Figure 1. After phase I, the cells are still allocated randomly along a multi-hop path and a packet experiences considerable delays waiting for the next transmission opportunity, which can be set back to the next slotframe (see Figure 1). B→A C→B E→A D→B F→E C→B D→B F→E B→A E→A B→A C→B F→E Slot offset C ha nn el o ffs et 0 1 2 3 0 Slotframe B E C FD Sink A Slotframe chunks [2, 3] [0, 1] Frequencies subset #1 Frequencies subset #2 0 1 0 1 Over-provisioned cell Schedule after phase I Schedule after phase II Timeslot Figure 1: 6TiSCH-CLX scheduling example End-to-end Delay Minimization: The second scheduling phase addresses the random al- location issue and ensures sink reachability from any leaf node within a single slotframe by actively utilizing the 6top sublayer alongside RPL topology information. The network is split into branches rooted at the sink. Each branch is assigned a unique subset of frequency channels and a full slotframe divided into ordered chunks based on the hop distance of the nodes (see Figure 1). Looking at the phase I random schedule, a packet generated at node C would always wait until the next slotframe to reach the sink, because the cell for link C → B occurs earlier in the slotframe than that of link B → A. In other words, the slot offsets are in descending order for uplink direction. Therefore, in phase II, nodes relocate their cells to align with the chunks boundaries and achieve an ascending sequence of transmission cells — a daisy-chain — starting at the leaf node. Adaptability to traffic patterns is also ensured through cell over-provisioning. 3 Analytical & Simulation Results To analyze the performance of the 6TiSCH-CLX framework, we develop a model of the network with the PRP and end-to-end packet delay as evaluation key metrics. To assess the impact of randomized cell scheduling the following two formulas can be used: E[dqn] = h ∑ i=2 i ( h−1 i ) 1 2h−1 , (1) fPRP = ( 1− pRc )D , (2) 3 / 5 Volume 080 (2021) A Novel Approach to Achieving End-to-End QoS for Avionic Applications where E[dqn] is the expected queuing delay, caused by the descending sequence of transmis- sion slot offsets along the path of h hops. Each pair of such links increases the queuing time by up to a whole slotframe, which is why the average number of such pairs in Equation 1 is counted based on the total number of hops h. fPRP — PRP distribution based on an assumption, that a packet is delivered successfully if no link along the path of D hops experiences more than the maximum retransmission count R collisions. To verify the advantages of the 6TiSCH-CLX framework, simulations based on the WAIC use cases from Table 1 are conducted in OMNeT++ with INET framework. Node layouts are shown in Figure 2 and 6TiSCH stack with MSF (6TiSCH-MSF) is used as a benchmark solution. 0.5m 1.5m Seat belt sensor Smoke alarm sensorLogical connection to S1 ... S1 Sn S1 1m 198 199 200195 196 197 1 2 3 4 5 6 Figure 2: Layout of the simulation scenarios with n sinks in the large-scale seat belt status (left) and one sink in the small-scale smoke alarm and humidity monitoring scenario (right) The end-to-end delay results for both small and large-scale scenarios are summarized in Fig- ure 3. For the smoke / humidity sensors, the impact of the distance to the sink (hop count) on latency is shown, while for the seat belt sensors, the effect of multiple sinks is investigated additionally. Overall, 6TiSCH-CLX achieves up to a double reduction in latency compared to 6TiSCH-MSF. While 6TiSCH-MSF fails to meet the application requirement of a 1s delay for more than 1 hop away from the sink, 6TiSCH-CLX supports up to 4 hops, after which queuing delays are becoming prevalent. The latency incurred by 6TiSCH-MSF is close to its analytical expectation, with slight deviation caused by the same queuing delays. The throughput and jitter in the seat belt scenario are visualized in Figure 4. With a single sink, 6TiSCH-CLX achieves slightly higher throughput and lower jitter than 6TiSCH-MSF, due to the lower and more consistent end-to-end delays for each packet. Acknowledgements: This publication is funded by the BMWi as part of the Retrofitbare Sensorsystem-Architektur für prädiktive Instandhaltung (ReSA) project, for ultra-resilient, retro- fittable wireless sensor networks for use in commercial aircraft, funding identifier 20X1721C. NetSys 2021 4 / 5 ECEASST 2 4 6 8 10 Hops 0 1 2 3 4 5 6 En d- to -e nd d el ay (s ) 6TiSCH-CLX 6TiSCH-MSF E[delay] MSF (analytical) CLX upper bound (best case) (a) Smoke / humidity sensors 1 2 3 4 5 Sinks 0 1 2 3 4 5 6 En d- to -e nd d el ay (s ) 6TiSCH-CLX 6TiSCH-MSF (b) Seat belt status Figure 3: End-to-end delay for small and large scale scenarios 1 2 3 4 5 Sinks 142 144 146 148 Th ro ug hp ut (B /s ) 6TiSCH-CLX 6TiSCH-MSF 1 2 3 4 5 Sinks 0.0 0.5 1.0 1.5 2.0 Jit te r ( s) 6TiSCH-CLX 6TiSCH-MSF Figure 4: Throughput and jitter in the seat belt scenario Bibliography [DALW15] S. Duquennoy, B. Al Nahas, O. Landsiedel, T. Watteyne. Orchestra: Robust mesh networks through autonomously scheduled TSCH. In Proceedings of the 13th ACM conference on embedded networked sensor systems. Pp. 337–350. 2015. [itu14] Operational and technical characteristics and protection criteria of radio altimeters utiliting the band 4200-4400 MHz. International Telecommunication Union (ITU) Recomendation ITU-R M.2059-0, 2014. [PAD+12] M. R. Palattella, N. Accettura, M. Dohler, L. A. Grieco, G. Boggia. Traffic aware scheduling algorithm for reliable low-power multi-hop IEEE 802.15. 4e networks. In 2012 IEEE 23rd International Symposium on Personal, Indoor and Mobile Radio Communications-(PIMRC). Pp. 327–332. 2012. [tsc12] IEEE Standard for Local and metropolitan area networks–Part 15.4: Low-Rate Wireless Personal Area Networks (LR-WPANs) Amendment 1: MAC sublayer. IEEE Std 802.15.4e-2012 (Amendment to IEEE Std 802.15.4-2011), 2012. 5 / 5 Volume 080 (2021) Introduction Cross-Layer Framework Analytical & Simulation Results