[PhD proposal] Dynamic management of real-time QoS in TSN networks

Title: Dynamic management of real-time QoS in TSN networks

Supervisor: Ye-Qiong Song (ye-qiong.song@loria.fr)

Hosting team: Simbiot (https://simbiot.loria.fr/)

Grant type : Competing for a doctoral contract of the University of Lorraine

Application deadline: 21/05/2020

 

Key words:

TSN (Time Sensitive Network), QoS (Quality of Service), Real Time, SDN (Software Defined Networking), Resource Allocation, Scheduling, Performance Evaluation.

 

Topic details:

TSN (Time Sensitive Network) is the latest important evolution of Ethernet. Pushed by the automotive industry as the future embedded network, TSN is also attracting the aerospace industry and the whole industry 4.0, thanks in particular to the numerous possibilities to ensure real-time data transfer, using different traffic shapers such as TAS (Time-Aware Shaper), CBS (Credit-Based Shaper) and BLS (Burst-Limiting Shaper) specified in IEEE802.1Qav, IEEE802.1Qbv.

SDN (Software Defined Networking) offers the possibility of monitoring and dynamic reconfiguration of network equipment (e.g. programmable switch/router) via the Openflow protocol or by P4 programming of both the control and data plan. This opens up extremely interesting prospects for reconfigure and reallocation of network resources (e.g. bandwidth, priority) to guarantee the QoS required for real-time flows in a switched network in general, and in TSN in particular.

Real-time applications are of various natures and requirements, the flows of data packets that a TSN network transmits must guarantee real-time QoS, and more specifically the bounded end-to-end delays, while optimizing the bandwidth utilization shared by these different flows. In addition, more and more real-time applications are dynamic in nature, requiring the TSN network to be able to admit new flows insert and existing flows withdrawal, while continuously guaranteeing real-time properties. This poses many challenging issues in dynamically managing QoS and network resource allocation.

This thesis aims to mainly develop methods for real-time QoS management and resource allocation of the TSN network. The idea is to exploit on the one hand the potentials of SDN, and on the other hand to contribute to the improvement of delay evaluation methods in the context of TSN, because the current delay upper bounds still present pessimism, leading to network resource overprovisioning.

Indeed, existing research work around TSN focuses on two topics: 1) optimal scheduling of time slots regulated by TAS [Dürr16], [Craciunas16], [Nayak18]; 2) evaluation of the worst-case response time of traffic regulated by CBS and BLS using different approaches such as the analysis of eligible intervals [Maxim17], [Cao18], network calculs [Zhao18], [Finzi19] and trajectorial analysis [Li17]. The remaining issues are the lack of joint consideration of both routing and scheduling problems, the lack of a study of the impact of time slot scheduling on lower priority traffic, and the pessimism of the delay bound evaluation, especially in a multi-hop TSN network, as well as a framework and executive mechanisms allowing dynamic admission of real-time flows, based on the various methods allowing to verify the real-time QoS guarantee.

In this PhD proposal, we will focus on the dynamic management of the real-time QoS in TSN. Specifically, we seek to jointly allocate time slots and paths in a TSN network in an optimal way. This allocation can be done through an admission control mechanism, i.e. each new request from flows (typically a flow of periodic data packets between a source and one or more destinations, with a constraint on the end-to-end delay across several TSN switches in the network) must be examined in order to predict if it can be accepted. In the case of acceptance, we need to allocate time slots and paths at each TSN switch. Since this is an NP-complete problem, the search for heuristic algorithms is preferred. The admission control algorithm will run on the SDN controller or on an application controller over SDN via a Northbound API (e.g. P4). An example of QoS management by an SDN-like controller is given in [Greff17]. Other work in this direction can be found in [Krishna16], [Durner15]. A first state of the art on SDN-based QoS management is also given in [Murat17]. It should be noted that there is not yet a programmable TSN switch by SDN (via commands Openflow) at the time of writing this PhD proposal. Nevertheless several recent works go in this direction [Greff17], [Schneider17], [Silva17]. The development can of course be firstly based on the emulation of switches (e.g. mininet, P4/BMv2).

In addition, depending on the candidate’s skills and interests, other avenues of contribution can be pursued, notably on reducing the pessimism of existing end-to-end delay limits, by exploiting other approaches that are more accurate but probably more complex (e.g. Forwarding analysis [Benammar18]) than that of network calculation and the eligible interval.

 

Candidate profile:

Master’s degree or equivalent in computer science or networking.

Good programming skills on c, java and python, and linux environment.

Knowledge and experience in one or more of the following topics will be appreciated: SDN, TSN, Optimization, Scheduling, Network calculus.

 

Bibliography:

[Benammar18] Nassima Benammar, Henri Bauer, Frédéric Ridouard, Pascal Richard, Timing analysis of AVB Ethernet network using the Forward end-to-end Delay Analysis. RTNS 2018: 223-233, Poitiers, France.

[Cao18] J. Cao, M. Ashjaei, P. J. L. Cuijpers, R. J. Bril, and J. J. Lukkien. An independent yet efficient analysis of bandwidth reservation for credit-based shaping in ethernet TSN. In 2018 IEEE World Conference on Factory Communication Systems (WFCS), pages 1–10, May 2018.

[Craciunas16] Silviu S. Craciunas, Ramon Serna Oliver, Martin Chmelik, and Wilfried Steiner, “Scheduling real-time communication in IEEE 802.1qbv time sensitive networks”. In Proceedings of the 24th International Conference on Real-Time Networks and Systems, RTNS ’16, pages 183–192, New York, NY, USA, 2016. ACM.

[Durner15] Raphael Durner, Andreas Blenk, Wolfgang Kellerer, Performance Study of Dynamic QoS Management for OpenFlow-enabled SDN Switches IEEE 23rd International Symposium on Quality of Service (IWQoS 2015), Portland, OR, USA, 15-16 June 2015.

[Dürr16] Frank Dürr and Naresh Ganesh Nayak, “No-wait packet scheduling for IEEE time-sensitive networks (TSN)”. In Proceedings of the 24th International Conference on Real-Time Networks and Systems, RTNS ’16, pages 203–212, New York, NY, USA, 2016. ACM.

[Finzi19] Anaïs Finzi, Ahlem Mifdaoui, Fabrice Frances, Emmanuel Lochin, Network Calculus-based Timing Analysis of AFDX networks incorporating multiple TSN/BLS traffic classes. CoRR abs/1905.00399 (2019).

[Greff17] F. Greff, Y.-Q. Song, L. Ciarletta, and A. Samama, 
“A Dynamic Flow Allocation Method to the Design of a Software-Defined Real-Time Mesh Network,” IEEE WFCS2017, May 31 – June 2, 2017, Trondheim, Norway.

[Krishna16] Hedi Krishna, Niels L. M. van Adrichem , Fernando A. Kuipers, Providing bandwidth guarantees with OpenFlow, IEEE Symposium on Communications and Vehicular Technologies (SCVT), 2016, Mons, Belgium.

[Li17] Xiaoting Li,Laurent George, Deterministic delay analysis of AVB switched Ethernet networks using an extended Trajectory Approach. Real-Time Systems 53(1): 121-186 (2017)

[Maxim17] Dorin Maxim and Ye-Qiong Song. Delay Analysis of AVB traffic in Time-Sensitive Networks (TSN). In RTNS 2017 – International Conference on Real-Time Networks and Systems, page 10, Grenoble, France, October 2017.

[Murat17] Murat Karakus and Arjan Durresi, Quality of Service (QoS) in Software Defined Networking (SDN) J. Netw. Comput. Appl., Vol. 80, pp200-218, 2017.

[Nayak18] Naresh Ganesh Nayak, Frank Dürr, Kurt Rothermel, Routing algorithms for IEEE802.1Qbv networks. SIGBED Review 15(3): 13-18 (2018).

[Schneider17] Ben Schneider, Alois Zoitl, Monika Wenger and Jan Olaf Blech, Evaluating Software-defined Networking for Deterministic Communication in Distributed Industrial Automation Systems IEEE ETFA2017, September 12-15, 2017, Limassol, Cyprus.

[Silva17] Luis Silva, Pedro Gonçalves, Ricardo Marau, Paulo Pedreiras and Luis Almeida, Extending Openflow with Flexible Time-Triggered Real-Time Communication Services IEEE ETFA2017, September 12-15, 2017, Limassol, Cyprus.

[Zhao18] L. Zhao, P. Pop, Z. Zheng, and Q. Li. Timing analysis of AVB traffic in TSN networks using network calculus. In 2018 IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS), pages 25–36, April 2018.

Logo d'Inria