Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs)

1. Project: IEEE P802.15 Working Group for Wireless Personal Area Networks (WPANs) Submission Title: [Time Slotted, Channel Hopping Field Experience] Date Submitted: [1 Sep, 2008] Source: [Kris Pister, Lance Doherty, Rick Enns, Kuor Hsin Chang, Clint Powell, José A. Gutierrez, Ludwig Winkel] Companies [Dust Networks, Freescale, Emerson, Siemens AG] Address [30695 Huntwood Avenue, Hayward, CA 94544 USA;890 N. McCarthy Blvd, Suite 120, Milpitas, CA 95035 USA; 8000 West Florissant Avenue St. Louis, Missouri 63136 USA; Siemensallee 74, Karlsruhe, Germany] Voice:[+1 (510) 400-2900, +1 (650) 327-9708, +1 (408) 904-2705, +1 (480) 413-5413, +1 (314) 553-2667,+49 (721) 595-6098] E-Mail:[kpister@dustnetworks.com, ldoherty@dustnetworks.com, enns@stanfordalumni.org, Kuor-Hsin.Chang@freescale.com, clinton.powell@freescale.com, Jose.Gutierrez@emerson.com,ludwig.winkel@siemens.com ] Re: [n/a] Abstract: [This document proposes extensions for IEEE802.15.4 MAC] Purpose: [This document is a response to the Call For Proposal, IEEE P802.15-08-373-01-0043] Notice:This document has been prepared to assist the IEEE P802.15. It is offered as a basis for discussion and is not binding on the contributing individual(s) or organization(s). The material in this document is subject to change in form and content after further study. The contributor(s) reserve(s) the right to add, amend or withdraw material contained herein. Release: The contributor acknowledges and accepts that this contribution becomes the property of IEEE and may be made publicly available by P802.15. Kris Pister et al.

2. Time Slotted, Channel Hopping (TSCH) Field Experience Kris Pister – UC Berkeley/Dust Networks Lance Doherty - Dust Networks Rick Enns - Consultant Kuor Hsin Chang - Freescale Clinton Powell - Freescale José A. Gutierrez – Emerson Ludwig Winkel – Siemens September, 2008 Kris Pister et al.

3. Overview • Presents empirical result from a multi-channel multi-hop Industrial Deployment • This is one of many working examples using TSCH technology. • Other examples at end of presentation if time permits • Measurement was taken for 26 days on all pair-wise channels in the network

4. Printing Factory Field Experience Topics • Network Topology & Location • Network Protocols • Time-Averaged Statistics • Time Series Data • Reliability in Uncertain Conditions • Summary of Results Kris Pister et al.

5. Network Topology • 44 Nodes • Gateway circled • 2.5 hop mean • Printing factory • 15,000 m2 • 3 floors • Concrete & steel Kris Pister et al.

6. Kris Pister et al.

7. Kris Pister et al.

8. Kris Pister et al.

9. Definitions • Each node has two parents • Childparent connection is a path • TDMA hopping over 16 channels of 2.402.48 GHz • Each path composed of 16 path-channels • Stability is the hop-by-hop packet success rate • We measured stability on all path-channels Parent Child Kris Pister et al.

10. Time-Averaged Stability for All Path-Channels Kris Pister et al.

11. Time-Averaged Stability per Path Kris Pister et al.

12. An Example Low-Stability Path • What does a path-channel look like? • How does it vary with time? • Let’s look at all 16 channels for a single path over time Kris Pister et al.

13. 26 Days: 2417 Path 2.48GHz 2.40GHz Kris Pister et al.

14. 17 47 56 44 Three Paths • Are paths geographically correlated? Kris Pister et al.

15. Three Paths - Stability Averaged over Time Kris Pister et al.

16. Three Paths - Stability Averaged over Time Kris Pister et al.

17. Three Paths - Channel 5 Over 26 Days Kris Pister et al.

18. Strategies to Overcome Variance • Path diversity • Have multiple parents for each node • Frequency diversity • Hop equally over all available channels • Time diversity • Link-layer ACKs and retries • Tolerate duplicates Kris Pister et al.

19. Reliability in the Midst of Variance • 44 nodes, 80B payload per packet • 33 packets per 15 min per node • 3.6 million packets, 17 lost • 99.9995% reliability over 26 days • All data secure and encrypted Kris Pister et al.

20. Expected Lifetime Assuming 2xAA batteries (3000mAh) Kris Pister et al.

21. Summary of Results • Average over time and frequency  good paths • Individual frequencies have periods of poor performance • Time-varying behavior unpredictable • Use network protocols to get mean behavior Kris Pister et al.

22. Conclusions • Industrial environments have varying channels • Low-power single-channel systems will have failures • Cannot predict performance • Average-case modeling software not applicable • Site surveys cannot capture behavior • Problems would be more severe with interference • Can appropriately overprovision to get reliability Kris Pister et al.

23. 2006: Cherry Point Refinery • Scope limited to Coker facility and support units spanning over 1200ft • No repeaters were needed to ensure connectivity • Electrical/Mechanical contractor installed per wired practices • >5 year life on C-cell 400m

24. 2006: Cherry Point Refinery • Scope limited to Coker facility and support units spanning over 1200ft • No repeaters were needed to ensure connectivity • Electrical/Mechanical contractor installed per wired practices • >5 year life on C-cell • >99.9% reliability 400m

25. Emerson MACTek Yokogawa Siemens Siemens ABB Honeywell Phoenix Contact Smar Endress+ Hauser Pepperl+ Fuchs Elpro Wireless HART interop demo, ISA 2006

26. Grane Platform, North Sea • 22 pressure sensors • 2 hour installation vs. 2 days Wireless Sensors

27. Shell Facility • Motor condition (vibration) monitoring • 200 temperature and vibration sensors • No line power due to hazardous location rules • Wiring in sensors would cause a 2 week delay in “first gas” HART Network 2 km 1 km

28. Pharmaceutical Process Monitoring • Temperature monitoring, latency tolerant, 100% of data required to avoid severe economic impact

29. Urban Infrastructure: Parking Monitoring • SF pilot 07: hundreds • LA pilot 08: 40,000

30. Reliable Performance in Harsh Environments Wireless Sensor • Steel mills • Chemical processing • Food production • Urban Pavement • Rail cars • Cracking towers • Pharmaceutical manufacturing • Desert fences • Northern coal facilities • Oil and gas facilities • … These and other factors conspire to define the difference between what works in the lab and what works in the real world! Steel mill scarfer