B-MAC Factored MAC protocol exposing control of sub-primitives

1. B-MACFactored MAC protocol exposing control of sub-primitives Joe Polastre <polastre@cs.berkeley.edu>

2. Outline • Factored MAC protocol design • Information sharing with higher layers • Control and reconfiguration of link protocol • Tradeoffs in link protocols • Feedback mechanism • Link Abstraction for Sensor Networks

3. Principles Reconfigurable MAC protocol Flexible control Hooks for sub-primitives Backoff/Timeouts Duty Cycle Acknowledgements Feedback to higher protocols Minimal implementation Minimal state Primary Goals Low Power Operation Effective Collision Avoidance Simple/Predicable Operation Small Code Size Tolerant to Changing RF/Networking Conditions Scalable to Large Number of Nodes Our implementation is on Mica2 motes with CC1000 B-MAC Design

4. B-MAC Link Protocol Interaction • Reconfiguration and control of link layer protocol parameters • Acknowledgements, Backoff/Timeouts, Power Management, Hidden Terminal Management • Ability to choose tradeoffs – “knobs” • Fairness, Latency, Energy Consumption, Reliability • Power consumption estimation through analytical and empirical models • Feedback to network protocols • Lifetime estimation • Mechanisms to achieve network protocols’ goals

5. Traditional monolithic protocol design Synchronized protocol with periodic listen periods “Black Box” design Designed for a general set of workloads User sets radio duty cycle SMAC takes care of the rest so you don’t have to Integrates higher layer functionality into link protocol T-MAC [van Dam and Langendoen, Sensys 2003] Reduces power consumption by returning to sleep if no traffic is detected at the beginning of a listen period sleep sleep listen listen sync sync sync sync Schedule 1 Schedule 2 Wei Ye, USC/ISI S-MACYe, Heidemann, and Estrin, INFOCOM 2002 Node 1 sleep sleep listen listen Node 2

6. Higher level communication scheduling Energy Cost = RX + TX + Listen Start by minimizing the listen cost Example of a typical low level protocol mechanism Periodically wake up, sample channel, sleep Properties Wakeup time fixed “Check Time” between wakeups variable Preamble length matches wakeup interval Overhear all data packets in cell Duty cycle depends on number of neighbors and cell traffic wakeup wakeup wakeup wakeup wakeup wakeup wakeup wakeup wakeup Low Power Listening (LPL) TX sleep sleep sleep Node 1 time RX sleep sleep sleep Node 2 time

7. Single hop data reporting application Higher sampling rate Higher traffic in a cell Higher duty cycle Optimize the check time to the traffic Application knows sample rate (packet generation rate) Effect of LPL Check Interval

8. Neighborhood Size affects amount of traffic in a cell Network protocols typically keep track of neighborhood size Bigger Neighborhood  More traffic Automatic Optimization Mac protocol can monitor channel Application vs network protocol vslink protocol Effect of Neighborhood Size

9. Factored vs Layered Protocols • Experimental Setup: • n nodes send as quickly as possible to saturate the channel • Factored link layer never worse than traditional approach • Often much better • Flexible configuration yields efficient: • Reliable transport (Acks) • Hidden Terminal support (RTS-CTS) topology

10. S-MAC RTS-CTS Fragmentation Support B-MAC Network protocol sends initial data packet with number of fragments pending Disable backoff & LPL for rest of fragments Measure energy consumption at C(bottleneck node) Minimizing power relieson controlling link layer primitives A E C B D Fragmentation SupportFactored vs Layered Protocol Sometimes the black boxis worse than the naïve approach

11. Assume a multihop packet is generated every 10 sec No queuing delay allowed Delay the packet S-MAC sleeps longer between listen period B-MAC increases the check interval and preamble length Tradeoffs: Latency for EnergyFactored vs Traditional Protocol S-MAC Default Configuration B-MAC Default Configuration

12. 10 node single hop network Increase transmission rate Deliver each packet within 10 sec Measure average power consumption per node As throughput increases B-MAC reduces check interval as traffic increases S-MAC uses optimal duty cycle Protocol overhead causes energy to increase linearly Tradeoffs: Throughput for EnergyFactored vs Layered Protocol topology

13. Surge Application • Run B-MAC in a real world application • 8 days/40000 data readings in deployment • Surge Multihop Data Collection includes: • Data reporting every 3 minutes • B-MAC check:sleep ratio: 1:100 • Jason Hill’s ReliableRoute • Reliability improvements to MintRoute • Turn on link layer acks in B-MAC • Add retransmissions • Power metering in the link protocol • Simple routing protocol optimization • Turn off long preambles when sending to the base station (one hop away) • Base station always on Network configuration images from Jason Hill

14. Duty cycle dependant on position in network Leaf nodes have least amount of traffic Middle nodes forward leaf node traffic Nodes 1 hop from base station benefit from reconfiguration 2.35% worst case node duty cycle Surge ApplicationNetwork power consumption of a factored link protocol Forwarded <10,000 packets 1 hop from base station • Forwarded ~35,000 (85%) packets • Duty cycle 75% higher • without optimization Leaf Nodes

15. Reliability 98.5% of all packets delivered Some nodes achieved an astounding 100% delivery …but communication links are volatile Retransmissions required After n retries, give up and pick a new parent Actual latency Follows expected latency from microbenchmark Retransmission delay Contention delay (infrequent) Tradeoffs: Latency vs ReliabilitySurge Application

16. B-MAC Analytical Model Worst case analysis Calculate Optimum LPL parameters Preamble length Evaluate Neighborhood size Sample rate Minimize Power consumption Verify Experimental operation matches analytical calculations Runtime Feedback mechanism Effect of turning knobs ti: LPL check interval r: Sample rate (packet generation rate) Knowledge from network protocols n: Neighborhood size Determine t: Fraction of each second spent on each operation E: Energy consumed by each operation per second Lifetime Model

17. Transmit Receive Lifetime Model

18. Lifetime Model • LPL Sampling • Sleep

19. The total energy, E, can be used to calculate the expected lifetime of the system Lifetime Model

20. Applications Aggregation RoutingServices “policy” Link Abstraction Physical Medium MAC Protocols Physical Layers “mechanism” Link Abstraction inWireless Sensor Networks • What have we learned from factored protocols? • Integration of routing and organization protocols with link protocols not necessary • Why do current WSN protocols integrate? • IP Nets: • Transport protocols set fragmenting, addressing type of service, retransmission etc • WSNs: Network protocols setting these parameters • Each hop has lots of volatility – keep state at every hop • Local per-link decisions to save power • IP model forces policy and mechanism to be integrated • Negates nice properties of IP abstraction:No protocol interchangeability, inefficient large implementations • A new abstraction must describe what the system is doing at each link • Dynamically change link protocol mechanism • Meeting point between link and routing layers • Separates link mechanisms from routing/network policies

21. Applications Aggregation RoutingServices “policy” Link Abstraction Physical Medium MAC Protocols Physical Layers “mechanism” Link Abstraction • B-MAC shows the need for bidirectional information flow with network protocols • Exposing control is critical for long lived operation • Enable link protocol interchangeability underneath optimized network protocols (routing, aggregation, organization, etc) • Smallest, most powerful primitives to execute higher level protocols efficiently • Proposed abstraction includes 3 things: • Control of link layer protocol parameters • Ability to choose tradeoffs – “knobs” • Power consumption feedback model • Next steps: • An RFC describing the abstraction in detail(this summer) • Implementation and use of abstraction in TinyOS with B-MAC/802.15.4

22. Conclusions • Coordination with higher protocols is essential for long lived operation • Feedback allows protocols to make informed decisions • Monolithic protocols can benefit from factoring—but reconfiguration may prove costly • Traditional abstraction at the network layer doesn’t fit sensor networks—need a new abstraction at the link layer

23. Backup Slides

24. Move protocol knowledge into the MAC implementation ARC Protocol Change the rate of originating traffic Allows route-through traffic to reach its destination Tradeoff everything else for fairness (very effective) Adaptive Backoff Change CSMA backoff depending on type of traffic Broadcast Multihop WooMACWoo and Culler, Mobicom 2001

25. Unified Network Protocol Framework (UNPF) Network organization protocol, Medium access control (MAC) protocol, and Routing protocol. Integrated network and link layers Write new UNPF implementations for each application -or- Trust the larger “black box” UNPFDing, Silvalingam, Kashyapa, and Chuan, Vehicular Technology Conference, 2003 Application Routing UNPF Routing, Organization, and MAC Organization Link Protocol Abstraction PHY

26. Energy-Aware TDMA-Based MAC Arisha, Youssef, Younis, IMPACCT 2002 Gateway node centrally manages sensor network clusters integrate organization protocol TRAMA: Collision Free Protocol Rajendran, Obraczka, Garcia-Luna-Aceves, SenSys 2003 Keep track of 2-hop neighbors, schedule traffic, eliminate collisions Adaptively change schedule as “inter-arrival” message time changes integrate organization protocol, assume traffic pattern Sift: Event Driven MAC Protocol Jamieson, Balakrishnan, Tay, MIT TR 2003 Small and fixed contention window to save energy Probabilistically pick an empty slot Moral: encourage protocol designers to expose configuration and tradeoffs (Sift even has a power consumption model) Other “Black Box” Protocols

27. Additional protocol overhead adds latency Choose applicable overhead ACK RTS-CTS (S-MAC) Synchronization B-MAC If multihop packet Wait >= 2 packet times to send next message Implicit reservation and hidden terminal support S-MAC Reserves channel at each hop 11 10 9 3 2 1 Tradeoffs: Latency vs ReliabilityFactored vs Traditional Protocol

28. Check time is 2.5ms If check interval is 250ms, duty cycle is 1% right? Wrong! Duty cycle is amortized cost of waking up energy consumption to sleep energy consumption Breakdown: Full power (e) 250ms Half power (b,d,f) 600ms Low power (c) 1500ms Equivalent to being at full power for 800ms Duty Cycle Calculation

29. Data transfer interface Establishes connections “Control” path ends at 802.2 Doesn’t pass down to underlying link protocols Doesn’t expose underlying capabilities of 802.3, 802.11, etc Instance specific code above 802.2 Linux 2.6.6 implementation Used as a programming interface, not an abstraction Heavyweight: 9008 lines of source, 28kB compiled Not an effective abstraction and not flexible Request Service Provider (802.2 LLC) Indication ServiceUser ServiceUser Response Confirm IEEE 802.2“Logical Link Control”IEEE Standard, 1998 802.2 Logical Link Control 802.1 Bridging / Convergence 802 .3 802 .11 802 .15 802 .16

30. Designed for low power, low data, high density rate wireless networks Services may interface with either 802.2 or directly with the MAC protocol 802.15.4 has lots of extra functionality that may never be used: Enable the development of a “lightweight 802.15.4 MAC” Use 15.4 hardware without using the 15.4 MAC protocol (eg: B-MAC, S-MAC, etc) IEEE 802.15.4“Low Rate Wireless Personal Area Networks”IEEE Standard, 2003 “Upper Layers” “Upper Layers” WSN Data Link Layer Abstraction x 802.2 802.1 802.15.4 MAC Other MAC 802.15.4 PHY 802.15.4 PHY