1. CS 577 / EE 537 Advanced Computer Networks Fall 2006 1 ExOR: Opportunistic Multi-Hop Routing for Wireless Networks Sanjit Biswas and Robert Morris M.I.T. Computer Science and Artificial Intelligence Laboratory Sigcomm 2005, Philadelphia Presented by Saurabh Gupta

2. CS 577 / EE 537 Advanced Computer Networks Fall 2006 2 Content  Background  ExOR: Overview  ExOR: Design  ExOR: Protocol  Evaluation  Summary

3. CS 577 / EE 537 Advanced Computer Networks Fall 2006 3 Wired Networks differ from Wireless Networks  Fixed nodes and fixed topology  Fixed Links between nodes  Independent Links between neighboring nodes  No great variance in link quality

4. CS 577 / EE 537 Advanced Computer Networks Fall 2006 4 Challenges for Wireless Routing  Capability constraints -Limited and varying range of transmission -Low Bandwidth Link -Common transmission medium shared by all nodes -Transmit power limitation  Mobility Dynamics -Topology changes frequently -Varying capacity  Transmission loss -Interference -Fading

5. CS 577 / EE 537 Advanced Computer Networks Fall 2006 5 Challenges for Wireless Routing (contd..)  Power Conservation  Asymmetry in forward and reverse links  Ease of snooping

6. CS 577 / EE 537 Advanced Computer Networks Fall 2006 6 ETX  The predicted number of data transmissions required to send a packet over a link  The ETX of a path is the sum of the ETX values of the links over that path  Examples: -ETX of a 3-hop route with perfect links is 3 -ETX of a 1-hop route with 50% loss is 2  Expected probability that a transmission is successfully received and acknowledged is d f x d r -d f is forward delivery ratio -d r is reverse delivery ratio  Each attempt to transmit a packet is a Bernoulli trial, so…

7. CS 577 / EE 537 Advanced Computer Networks Fall 2006 7 Traditional Routing  Abstract radio to look like a wired link  Identify a route  Packets get forwarded on fixed path  Retried on failures  Looks like a circuit switched network packet src AB dst C D

8. CS 577 / EE 537 Advanced Computer Networks Fall 2006 8 Cooperative Diversity  Broadcast transmission over Wireless Links involves probabilistic delivery  Sends information through multiple relays, concurrently  Destination chooses best of many relayed signals, or combine information from multiple signals  Requires radios capable of simultaneous, synchronized repeating of signals or additional radio channels for each relay 1234561 23635 1 42345612456 src AB dst C

9. CS 577 / EE 537 Advanced Computer Networks Fall 2006 9  Background  ExOR: Overview  ExOR: Design  ExOR: Protocol  Evaluation  Summary

10. CS 577 / EE 537 Advanced Computer Networks Fall 2006 10 ExOR (Extremely Opportunistic Routing)  Integrated link/network-layer diversity routing technique  Realizes some of the gains of cooperative diversity on standard radio hardware, such as 802.11  Uses broadcasts for large unicast transfers in multi-hop wireless networks  Uses Delayed Forwarding -does not make the forwarding decision until after reception -only the “best” receiver of each packet forwards it -avoids duplication  Operates on batches of packets, to reduce communication cost of agreement

11. CS 577 / EE 537 Advanced Computer Networks Fall 2006 11 packet src AB dst C packet ExOR: Basic Functionality  Decide which nodes receive broadcasts  Decide who forwards, after reception  Node closest to the destination forwards

12. CS 577 / EE 537 Advanced Computer Networks Fall 2006 12 packet src AB dst C packet ExOR: Basic Functionality (contd…)

13. CS 577 / EE 537 Advanced Computer Networks Fall 2006 13 N1N3 N5 N7N6N2N4N8SD Traditional Path  Traditional routing must compromise between hops to choose ones that are long enough to make good progress but short enough for low loss rate  With ExOR each transmission may have more independent chances of being received and forwarded  It takes advantage of transmissions that reach unexpectedly far, or fall unexpectedly short How ExOR might provide more throughput

14. CS 577 / EE 537 Advanced Computer Networks Fall 2006 14  Traditional routing: 1 / 0.25 + 1 = 5 transmissions  ExOR: 1 /(1 – (1 – 0.25) 4 ) + 1 = 2.5 transmissions  Assumes independent losses N1 srcdst N2 N3 N4 25% 100% How ExOR might provide more throughput (contd..)

15. CS 577 / EE 537 Advanced Computer Networks Fall 2006 15  Background  ExOR: Overview  ExOR: Design  ExOR: Protocol  Evaluation  Summary

16. CS 577 / EE 537 Advanced Computer Networks Fall 2006 16 ExOR: Design Challenges  Agreement amongst the nodes on which sub-set of them received each packet  Algorithm to decide, from amongst the receiving nodes, the node “closest” to the ultimate destination that forwards the packet –requires a metric reflecting the likely cost of moving a packet from any node to the destination  Algorithm to choose only the most useful nodes as participants  Avoid simultaneous transmissions by different nodes to minimize collisions

17. CS 577 / EE 537 Advanced Computer Networks Fall 2006 17 Node States for each batch it participates in  Packet Buffer -stores successfully received packets, according to the batch numbers  Local Forwarder List -is the copy of prioritized list of nodes, copied from one of the packets -for a given batch, all nodes use same forwarder list as specified by the sender  Forwarding Timer -time at which the node predicts that it should forward packets -timer set far enough ahead to give higher priority nodes enough time to send -adjusted when packets from other nodes heard

18. CS 577 / EE 537 Advanced Computer Networks Fall 2006 18 Node States for each batch it participates in (contd…)  Transmission Tracker -records measured rate of current sending node and expected number of packets left to send -used by node to adjust forwarding timer -adapts to competing traffic  Batch map -indicates, for each packet in a batch, the highest-priority node to have received a copy of that packet

19. CS 577 / EE 537 Advanced Computer Networks Fall 2006 19 ExOR: Packet Format -HdrLen & PayloadLen indicate size of ExOR header and payload respectively -PktNum is current packet’s offset in the batch, corresponding to the current batch-map entry -FragSz is size of currently sending node’s fragment (in packets) -FragNum is current packet’s offset within the fragment -FwdListSise is is number of forwarders in list -ForwarderNum is current sender’s offset within the list -Forwarder List is copy of sender’s local forwarder list -Batch Map is copy of sending node’s batch map, where each entry is an index into Forwarder List

20. CS 577 / EE 537 Advanced Computer Networks Fall 2006 20  Background  ExOR: Overview  ExOR: Design  ExOR: Protocol  Evaluation  Summary

21. CS 577 / EE 537 Advanced Computer Networks Fall 2006 21  How often should ExOR run? -Per packet is expensive -Use batches  Who should participate in the forwarding? -Too many participants cause large overhead  When should each participant forward? -Avoid simultaneous transmission  What should each participant forward? -Avoid duplicate transmission  How and When does the process complete? -Identify the convergence of the algorithm ExOR: Protocol

22. CS 577 / EE 537 Advanced Computer Networks Fall 2006 22 Who should participate?  The source chooses the participants (forwarder list) using ETX-like metric -Only considers forward delivery rate  The source runs a simulation and selects only the nodes which transmit at least 10% of the total transmission in a batch -A background process collects ETX information via periodic link-state flooding

23. CS 577 / EE 537 Advanced Computer Networks Fall 2006 23 When should each participant forward?  Forwarders are prioritized by ETX-like metric to the destination  Receiving nodes buffer successfully received packets till the end of the batch  The highest priority forwarder transmits from its buffer when the batch ends -These transmissions are called the node’s fragment of the batch  The remaining forwarders transmit in prioritized order  Question: How does each forwarder know it is its turn to transmit -Assume other higher priority nodes send for five packet durations if not hearing anything from them

24. CS 577 / EE 537 Advanced Computer Networks Fall 2006 24 What should each participant forward?  Packets it receives yet not received by higher priority forwarders  Each packet includes a copy of the sender’s batch map, containing the sender’s best guess of the highest priority node to have received each packet in the batch  Question: How does a node know the set of packets received by higher priority nodes? -Using batch map

25. CS 577 / EE 537 Advanced Computer Networks Fall 2006 25 How and When does the process complete?  If a node’s batch map indicates that over 90% of the batch has been received by higher priority nodes, the node sends nothing when its turn comes  When ultimate destination’s turn comes to send, it transmits 10 packets including only its batch map and no data  Question: How is the remaining 10% data delivered? -Using traditional routing

26. CS 577 / EE 537 Advanced Computer Networks Fall 2006 26 Gossip Mechanism for Reliable summaries  Repeat summaries in every data packet  Cumulative: what all previous nodes rx’d N0 N1 N2 N3 1 st round Tx: 1, 2, 3, 4, 5, 6, 7, 8 Forwarder list: N3(dst), N2, N1, N0 (src) Rx:2,5,8 Rx: 2,4 Rx: 1,2,7,8 Batch map: 03030000 Batch map: 03032002 Batch map: 13032012 Tx: 5,8 Tx: 1,7 2 nd round Tx: 3,6 Batch map: 13032012

27. CS 577 / EE 537 Advanced Computer Networks Fall 2006 27 Transmission Timeline for an ExOR transfer N24 not able to listen to N5. N8 does not send N17 might have missed some batch-maps

28. CS 577 / EE 537 Advanced Computer Networks Fall 2006 28  Background  ExOR: Overview  ExOR: Design  ExOR: Protocol  Evaluation  Summary

29. CS 577 / EE 537 Advanced Computer Networks Fall 2006 29 1 kilometer 65 Roofnet node pairs

30. CS 577 / EE 537 Advanced Computer Networks Fall 2006 30 Evaluation Details  65 Node pairs  1.0MByte file transfer  1 Mbit/s 802.11 bit rate  1 KByte packets  Batch size: 100 packets  ExOR Header: 44-114 bytes Traditional RoutingExOR 802.11 unicast with link-level retransmissions Hop-by-hop batching UDP, sending as MAC allows 802.11 broadcasts 100 packet batch size  Reported values are median of nine experimental runs, to reduce effect from other user traffic and other sources

31. CS 577 / EE 537 Advanced Computer Networks Fall 2006 31 Throughput (Kbits/sec) 1.0 0.8 0.6 0.4 0.2 0 0200400600800 Cumulative Fraction of Node Pairs ExOR Traditional ExOR: 2x Improvement in throughput Median throughputs: 240 Kbits/sec for ExOR, 121 Kbits/sec for Traditional Figure 8: The distribution of throughputs of ExOR and traditional routing between the 65 node pairs. The plots shows the median throughput achieved for each pair over nine experimental runs.

32. CS 577 / EE 537 Advanced Computer Networks Fall 2006 32 25 Highest throughput pairs Node Pair Throughput (Kbits/sec) 0 200 400 600 800 1000 ExOR Traditional Routing 1 Traditional Hop 1.14x 2 Traditional Hops 1.7x 3 Traditional Hops 2.3x  For single hop pairs ExOR provides the advantage of lower probability of source resending packets, as there’s higher probability of source receiving the destination’s 10 batch-map packets Figure 9: The 25 highest throughput pairs, sorted by traditional routing throughput. The bars show each pair's median throughput, and the error bars show the lowest and highest of the nine experiments.

33. CS 577 / EE 537 Advanced Computer Networks Fall 2006 33 25 Lowest throughput pairs Node Pair 4 Traditional Hops 3.3x Longer Routes Throughput (Kbits/sec) 0 200 400 600 800 1000 ExOR Traditional Routing Figure 10: The 25 lowest throughput pairs. The bars show each pair's median throughput, and the error bars show the lowest and the highest of the nine experiments. ExOR outperforms traditional routing by a factor of two or more.  As number of node pairs increases along a route, the likelihood of increased choice of forwarding nodes and multiple ways to ‘gossip’ back batch-maps, increases  With greater routing length ExOR is able to take advantage of asymmetric links also

34. CS 577 / EE 537 Advanced Computer Networks Fall 2006 34 Retransmissions affected by selection of hops Traditional routing has to select the ‘shortest’ path which results in compromise on selecting drop probability, thus increasing the number of transmissions ExOR has no limitations on number of nodes, from the forwarder list, that can forward the packet. Hence it uses both nodes closer to source and nodes closer to destination, irrespective of their drop probability Figure 11: The number of transmissions made by each node during a 1000-packet transfer from N5 to N24. The X axis indicates the sender's ETX metric to N24. The Y axis indicates the number of packet transmissions that node performs. Bars higher than 1000 indicate nodes that had to re-send packets due to losses.

35. CS 577 / EE 537 Advanced Computer Networks Fall 2006 35 ExOR moves packets farther Figure 12: Distance traveled towards N24 in ETX space by each transmission. The X axis indicates the di®erence in ETX metric between the sending and receiving nodes; the receiver is the next hop for traditional routing, and the highest-priority receiving node for ExOR. The Y axis indicates the number of transmissions that travel the corresponding distance. Packets with zero progress are not received by the next hop (for traditional routing) or by any higher-priority node (for ExOR). Max. distance traveled by hops in traditional routing Distance traveled by transmissions in ExOR Big chunk of transmission, in traditional routing, takes place over shorter distances Number of packets carried over individual long distance links is small But cumulative transmission is substantial

36. CS 577 / EE 537 Advanced Computer Networks Fall 2006 36 ExOR moves packets farther  Delivery Probability decreases with distance  ExOR average: 422 meters/transmission  Traditional Routing average: 205 meters/tx Fraction of Transmissions 0 0.1 0.2 0.6 ExOR Traditional Routing 01002003004005006007008009001000 Distance (meters) 25% of ExOR transmissions 58% of Traditional Routing transmissions

37. CS 577 / EE 537 Advanced Computer Networks Fall 2006 37 ExOR uses links in parallel Traditional Routing 3 forwarders 4 links ExOR 7 forwarders 18 links

38. CS 577 / EE 537 Advanced Computer Networks Fall 2006 38 Batch Size  ExOr header grows with the batch size  Large batches work well for low-throughput pairs due to redundant batch map transmissions  Small batches work well for high throughput pairs due to lower header overhead

39. CS 577 / EE 537 Advanced Computer Networks Fall 2006 39  Background  ExOR: Overview  ExOR: Design  ExOR: Protocol  Evaluation  Summary

40. CS 577 / EE 537 Advanced Computer Networks Fall 2006 40 Summary  Integrated routing and MAC protocol for Multi-Hop Wireless Networks  Uses Delayed forwarding mechanism, whereby the forwarding decision is made only after reception of packets  Takes advantage of the probabilistic nature of wireless broadcast transmissions, and does not hide it  ExOR does not require separate control signals for agreement  Forwarding based on reception of the given data packets and not on signal strength measurements or previous control / data packets  Less dependent upon channel stability  “Gossip” mechanism reduces the likelihood of duplicate transmissions

41. CS 577 / EE 537 Advanced Computer Networks Fall 2006 41 Summary (contd..)  Fewer transmissions of each packet  Utilizes long asymmetric links  Increases total network capacity  Increases individual connection throughput (approx. 2x)  Requires link-state graphs  Introduces overhead in form of “batch info”  Difficult to scale over large, dense networks

42. CS 577 / EE 537 Advanced Computer Networks Fall 2006 42 Acknowledgements Many sketches, animated-diagrams, as well as some text have been sourced from the following materials- Course material on “Net Centric Systems” taught at TECHNISCHE UNIVERSITÄT DARMSTADT Presentation on “A High Throughput Route-Metric for Multi-Hop Wireless Routing” by Eric Rozner of University of Texas, Austin Presentation on “ExOR: Opportunistic Multi-Hop Routing for Wireless Networks”, by Sanjit Biswas and Robert Morris at Siggcomm “ExOR: Opportunistic Multi-Hop Routing for Wireless Networks” - Sanjit Biswas and Robert Morris Presentation on “ExOR: Opportunistic Multi-Hop Routing for Wireless Networks”, by Avijit of University of California, Santa Barbara Presentation on “ExOR: Opportunistic Multi-Hop Routing for Wireless Networks”, by Yu Sun of University of Texas, Austin Presentation on “ExOR: Opportunistic Multi-Hop Routing for Wireless Networks”, by Gaurav Gupta, University of Southern California Presentation on “ExOR: Opportunistic Multi-Hop Routing for Wireless Networks”, by Ao-Jan Su, Northwestern University

43. CS 577 / EE 537 Advanced Computer Networks Fall 2006 43 Questions? Thank you!!