Chair of Network Architectures and Services Department of Informatics Technical University of Munich Evaluation of Online Schedule Synthesis Algorithms for Time-based Scheduled Time Sensitive Networks Intermediate talk for the Master’s Thesis by Alexander Mildner advised by Max Helm, Benedikt Jaeger, Dr. Marcel Wagner (Intel), Hector Blanco Alcaine (Intel) Monday 30 th September, 2019 Chair of Network Architectures and Services Department of Informatics Technical University of Munich
Outline • Introduction to Time Sensitive Networking • Thesis Objectives • Approach • Network Calculus Model for WCD Analysis in TSN Networks • Building a TSN emulation environment based on Mininet • GCL Synthesis Algorithms • Testbed Evaluations • Further Steps and Future Outlook A. Mildner — TSN 2
Introduction to Time Sensitive Networking • Set of IEEE standards and additions to the IEEE 802.3 Ethernet standard, for providing deterministic services and meet the requirements for bounded latency packet transmission on Layer 2 • IEEE Time Sensitive Networking Task Group, former Audio Video Bridging (AVB) Task Group • High interest on TSN in the following industry sectors: • Factroy Automation (Industry 4.0) • Automotive • Aerospache/Avionic • Autonomous Driving • etc. • The Standarization process is still ongoing, some important standards are not yet finished and published A. Mildner — TSN 3
Introduction to Time Sensitive Networking IEEE TSN Standards Figure 1: TSN components [2] A. Mildner — TSN 4
Thesis Objectives Problem Statement • Currently the TSN standardization is lacking a proper way for dynamically (re-)configure TSN-Networks • The majority of the shown TSN-Demonstrations (e.g. on fairs) are statically configured and only work with prior knowledge about the complete network • The scheduling problem introduced by IEEE 802.1Qbv (Time-based Scheduling) can be- come non-trival to solve for large Networks • But: There have been recent research efforts, in order to tackle the dynamic Schedule Synthesis Goal of this Thesis: Compare and evaluate different approaches for dynamic Schedule Synthe- sis for time-based scheduled TSN-Networks A. Mildner — TSN 5
Thesis Objectives • Objective 1: Implement Network Calculus Model for determining Worst Case Delay (WCD) of flows in Time Sensitive Networks with static Schedule Configuration • Objective 2: Implement an emulation environment based on mininet for time-based scheduled TSN Networks • Objective 3: Implement, validate and optimize Schedule Synthesis Algorithms for time-based scheduled TSN Networks • Objective 4: Evaluation and comparison the different Synthesis Algoritms according to performance, execution time and validity of the generated Schedules • Objective 5 (Optional): Create a testbed/demonstrator utilizing the evaluated online mechnaisms for dynamic sched- ule generation and adaption on real TSN-capable Hardware A. Mildner — TSN 6
Approach IEEE 802.1Qbv - Time Aware Scheduler GCL Q7 Q6 Q5 Q4 Q3 Q2 Q1 Q0 T 1 : 1000 0000 T 2 : 0111 1111 T 3 : 0111 1111 T 4 : 0111 1111 T 5 : 0111 1111 T GCL Gate Gate Gate Gate Gate Gate Gate Gate Transmission Selection Figure 2: Time-based Scheduler (GCL: Gate Control List, T GCL : Cycle Time of the schedule) A. Mildner — TSN 7
Approach IEEE 802.1Qbv - Time Aware Scheduler GCL Q7 Q6 Q5 Q4 Q3 Q2 Q1 Q0 T 1 : 1000 0000 T 2 : 0111 1111 T 3 : 0111 1111 T 4 : 0111 1111 T 5 : 0111 1111 T GCL Gate Gate Gate Gate Gate Gate Gate Gate Transmission Selection Figure 3: Time-based Scheduler (GCL: Gate Control List, T GCL : Cycle Time of the schedule) A. Mildner — TSN 8
Approach Example TSN Network CUC CNC ES1 ES3 Flow 1 GCL GCL GCL SW1 SW2 GCL GCL GCL Flow 2 ES2 ES4 Figure 4: Template A. Mildner — TSN 9
Approach Network Calculus Model for WCD Analysis in TSN Networks • L. Zaho et. al proposed in [4] a Network Calculus model for determining the Worst Case Delay of flows in time-based scheduled TSN Networks • Works on statically configured time-based scheduled TSN Networks (GCLs given) • Inputs → network topology, GCLs, Flow information, Flow of Interest (FOI) • Output → End-to-End WCD of the FOI • Implemented Parameter Calculation for the resulting service curves ( β ) for each port on the FOIs Path in Python • A lot of work tended to be optimal preparation of the input data for calculation of Overlap- ping scenarios (more on next slide) • Opens: • Feed parameters into DiscoDNC 1 Framework for reliable WCD calculation • Validate our implementation using the in [4] presented evaluation results 1 https://disco.cs.uni-kl.de/index.php/projects/disco-dnc A. Mildner — TSN 10
Approach Building a TSN emulation environment based on Mininet • Nice to Have: A flexible Network Emulator that can do time-based scheduling according to the standard • Idea: Use recently added TAPRIO 2 kernel net-scheduler module to add time-based sched- uling capabilities to mininet 3 • TAPRIO can be easily configured as any other Queuing Dicipline (qdisc) using the iproute2 tool tc (example next slide) • Problem: • mininet uses veth (virtual ethernet) interfaces • veth interfaces implement no Transmit (TX) Queues • TAPRIO requires TX Queues to work • Opens: • Adapt veth implementation to support multiple TX queues and thus supports TAPRIO • Create a test framework on running automated evaluations for the Network Calculus Model and the GCL Synthesis Algorithms on various netwrok topologies 2 http://man7.org/linux/man-pages/man8/tc-taprio.8.html 3 http://mininet.org/ A. Mildner — TSN 11
Approach Overlapping Scenarios in GCLs A. Mildner — TSN 12
Approach GCL Synthesis Algorithms • Two main Related Works on GCL Synthesis Algorithms: • Algorithm 1: R. Oliver et. al proposed in [3] a GCL Synthesis algorithm based on Array Theory Encoding ( T A ) for Satisfiability Modulo Theory (SMT) • Algorithm 2: S. Craciunas et. al proposed in [1] a GCL Synthesis Algorithm based on Integer Linear Programming (ILP) for SMT • Some other interesting approaches: No-wait Packet Scheduling Problem (NW-PSP), sim- ple heuristic approaches • Opens: • Implement and Evaluate Algorithm 1 and 2 using z3 4 SMT/OMT solver for Python • Implement and Evaluate a simple heuristic schedule synthesis algorithm 4 https://github.com/Z3Prover/z3 A. Mildner — TSN 13
Approach Testbed Evaluations • Idea: Show off the functionality of a dynamically configurable time-based scheduled TSN Network on real Hardware • Most of the required software ( sender, receiver, visualization/control tool ) is already implemented • Once the GCL Synthesis Algorithms have been implemented, we can setup a demonstrator • User can add/remove flows with certain properties to the TSN-Network and network will automatically re-configure itself and shows off acheived end-to-end packet latencies for the FOI • Open: Finish implementation and setup of the demonstrator, once all missing pieces are done A. Mildner — TSN 14
Further Steps and Future Outlook Further Steps and Future Outlook • Network Calculus Model implementation almost finished, first evaluation results expected soon • Facing some problems regarding enabling TAPRIO on mininet (How can we implement TX Queues on veth interfaces ?) • For the next major phase of this thesis, we will focus on the implementation, evaluation and possible optimizations of the GCL Synthesis Algorithms • If we have enough time, we will create a demonstrator, which shows the dynamic schedule adaption using real TSN-capable Hardware Thanks for your Attention ! A. Mildner — TSN 15
Backup TAPRIO configuration example tc qdisc add dev IFACE parent root handle 100 taprio num_tc 3 # Number of traffic Classes map 2 2 1 0 2 2 2 2 2 2 2 2 2 2 2 2 # Map Traffic Class -> SKB Priority queues 1@0 1@1 2@2 # Map Traffic Class -> HW Queue base-time 10000000 # Start Time # 1 st Schedule Entry sched-entry S 03 300000 # 2 nd Schedule Entry sched-entry S 02 300000 # 3 rd Schedule Entry sched-entry S 06 400000 clockid CLOCK_TAI # Clock Source to use (Reference Clock) A. Mildner — TSN 16
Bibliography [1] S. S. Craciunas, R. S. Oliver, M. Chmelík, and W. 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. [2] J. Farkas. Introduction to IEEE 802.1 - Focus on the Time-Sensitive Networking Task Group, 2017. http://www.ieee802.org/1/files/public/docs2017/tsn-farkas-intro-0517-v01.pdf. [3] R. Serna Oliver, S. S. Craciunas, and W. Steiner. Ieee 802.1qbv gate control list synthesis using array theory encoding. In 2018 IEEE Real-Time and Embedded Technology and Applications Symposium (RTAS) , pages 13–24, April 2018. [4] L. Zhao, P . Pop, and S. S. Craciunas. Worst-case latency analysis for ieee 802.1qbv time sensitive networks using network calculus. IEEE Access , 6:41803–41815, 2018. A. Mildner — TSN 17
Recommend
More recommend
