Reliable Multihop Transfer on Wireless Sensor Networks Rodrigo - PowerPoint PPT Presentation

Reliable Multihop Transfer on Wireless Sensor Networks Rodrigo Fonseca, Sukun Kim, David Culler University of California Berkeley IEEE SECON – October 2004

Motivation � Some sensornet applications require 100% reliability over multiple hops � Structure monitoring � Logging (development, deployment) � Auditing � This has proven to be non- trivial IEEE SECON – October 2004

Challenges � Wireless communications � Low power radios � Asymmetric, changing links � Interference, etc. � Resource constrained hardware � Memory, computational power, energy IEEE SECON – October 2004

Problem Scope � Design options for achieveing high reliability over multiple hops � Traffic pattern: � One destination, large data (in comparison to pkt) � Focus on convergence and point-to-point � Assumption: � Routing layer provides a path, or set of paths to the destination IEEE SECON – October 2004

Design Options � Only a fraction of the transmissions goes through any given link � We can improve reliability by increasing � Number of packets injected � Probability of success � Redundancy � Retransmissions -- End to End, Link Level, Both � Erasure coding � Probability of Success � Path selection, Alternate paths � Congestion Control IEEE SECON – October 2004

Outline � Introduction � Alternatives for Reliability � Alternatives for Reliability � Retransmission � Erasure Coding � Alternate Routes � Experimental Results � Conclusions IEEE SECON – October 2004

End-to-End Retransmission � Probability of success over multiple hops decreases rapidly � Wasted effort � Fail at hop n, n-1 wasted transmissions � E2E path may not exist at all times � Reverse path may not exist � Although for 100% reliability source must receive some signal from destination IEEE SECON – October 2004

Link Level Retransmission � Found to be very efficient in increasing reliability � Effect of boosting each link success probability � Local repair IEEE SECON – October 2004

Link Level Retransmission � Testbed experiment, 5 hops, avg link quality 77% 1 0.9 0.8 0.7 Success Rate 0.6 0.5 0.4 0.3 0.2 0.1 Empirical Theoretical 0 0 1 2 3 4 5 Maximum number of retransmission IEEE SECON – October 2004

Link Level Retransmission � Found to be very efficient in increasing reliability � Effect of boosting each link success probability � Local repair � However, still fails to reach 100% reliability � Bursty loss pattern � Cost of achieving even higher reliability may become very high IEEE SECON – October 2004

Erasure Codes � What if we can tolerate the loss of a few percent of the packets? � Transmit redundant information � Erasure codes allow M out of M+N packets to be recovered � Fraction of redundancy is called rate of code � There are rateless codes, which can produce unlimited redundancy, but may be expensive IEEE SECON – October 2004

Erasure Codes Encoding Channel Decoding M M 8 msgs 8 original msgs N N’ 21 code words = 8 code words IEEE SECON – October 2004

Systematic Codes Benefit: if receiver has codes containing original messages • Encoding, Decoding are faster • Even if receiver get less than 8 packets, we don’t lose every message IEEE SECON – October 2004

Implementation on TinyOS � We use systematic codes � No memory overhead for encoding or decoding � Codewords generated on the fly � Reception can stop once M pkts received � Real time operation on Mica2 motes � Available for TinyOS IEEE SECON – October 2004

But Losses are Bursty... � If we loose more than N-M packets, can’t recover the entire data � Codes introduce a fixed redundancy overhead � So, depending on the loss process � Waste bandwidth on all packets � Not effective when needed... IEEE SECON – October 2004

Alternative Routes Find Alternative Route: a form of ‘spatial retransmission’ But this may get tricky if we get a lot of failures Get k best candidates for the next hop from routing layer, and try from the best IEEE SECON – October 2004

Alternative Routes � Dynamic alternative route selection � Provides immediate reaction to failed route � We change the routing layer to provide possible next hops, instead of one � Successively try alternatives � May still drop packet if no possible route works IEEE SECON – October 2004

Outline � Introduction � Alternatives for Reliability � Retransmission � Erasure Coding � Alternate Routes � Experimental Results � Experimental Results � Conclusions IEEE SECON – October 2004

Experimental Setup � Point-to-point routing � Beacon Vector Routing used to provide routes, remained stable � Soda Hall Testbed, 78 motes � 1 pair of nodes at a time, 300 packets @ 1/s � Results shown: � 1 pair of nodes, 300 packets @ 1/s � Average route 5 hops � Other pairs similar results IEEE SECON – October 2004

Testbed Source Destination IEEE SECON – October 2004

Metrics � We are mainly concerned with two metrics: � Reliability � Fraction of application data packets that are received by the destination � ‘Work’ � Number of transmissions per successfully received packet, per hop � Ideally, 1 transmission per hop per message IEEE SECON – October 2004

Reliability 1.2 1 Success Rate 0.8 Redundancy Erasure code 0.6 redundancy 0 0.4 2 4 0.2 8 0 0 1 2 3 4 5 5+AR Maximum number of retransmissions IEEE SECON – October 2004

Work per packet, per hop 4 3.5 Work (per packet, per hop) 3 0 2.5 1 2 2 4 1.5 8 Redundancy 1 0.5 AR means alternate route is used 0 0 1 2 3 4 5 5+AR Maximum number of retransmissions IEEE SECON – October 2004

Reliability versus work 1 8 0 0.9 0 0.8 3 Retransmissions 0.7 2 Redundancy 1 0 0 0.6 Reliability 1 0.5 2 5 0.4 5+AR 0.3 8 0 1 2 3 4 5 0.2 0.1 0 1 1.5 2 2.5 3 3.5 4 Work (packets per received data packet, per hop) IEEE SECON – October 2004

Finding the best combination Given a threshold reliability requirement, what is the retransmission /redundancy combination that has the smallest overhead? Work IEEE SECON – October 2004

Conclusions and Future Work � Main Contributions � We evaluated different combinations of options for multihop reliability � Implementation of real time erasure coding on TinyOS � Combination of options yields best results � Erasure coding allows packet drops � Alternate route makes the loss process more amenable to erasure coding � Important for routing layer to quickly detect and route around failure � Future work � Throrough characterization of loss patterns in other settings � Experiments with different routing algorithms IEEE SECON – October 2004