Presentation is loading. Please wait.

Presentation is loading. Please wait.

7: Wireless Ad Hoc Networks7-1 Chapter 7 Wireless Ad Hoc Networks.

Similar presentations


Presentation on theme: "7: Wireless Ad Hoc Networks7-1 Chapter 7 Wireless Ad Hoc Networks."— Presentation transcript:

1 7: Wireless Ad Hoc Networks7-1 Chapter 7 Wireless Ad Hoc Networks

2 7: Wireless Ad Hoc Networks7-2 r Definitions: m An ad-hoc network is one that comes together as needed, not necessarily with any assistance from the existing Internet infrastructure m Instant infrastructure m A MANET is a collection of mobile platforms or nodes where each node is free to move about arbitrarily m A MANET: distributed, possibly mobile, wireless, multihop network that operates without the benefit of any existing infrastructure (infrastructure-less), except the nodes themselves What is an Ad Hoc Network?

3 7: Wireless Ad Hoc Networks7-3 Mobile Ad Hoc Networks r May need to traverse multiple links to reach a destination

4 7: Wireless Ad Hoc Networks7-4 Mobile Ad Hoc Networks (MANET) r Mobility causes route changes

5 7: Wireless Ad Hoc Networks7-5 Why Ad Hoc Networks ? r Ease of deployment r Speed of deployment r Decreased dependence on infrastructure

6 7: Wireless Ad Hoc Networks7-6 Fundamental Challenges It is better to know some of the questions than all of the answers. — James Thurber ( )

7 7: Wireless Ad Hoc Networks Energy Efficiency r No infrastructure means must rely on batteries (or, in general, limited energy resources) r Possible solutions m Selectively sending nodes into a sleep mode m Using transmitters with variable power (the Power Control problem) m Using energy-efficient paths m Using cooperative techniques (still relatively new)

8 7: Wireless Ad Hoc Networks Mobility r Mobility-induced route changes r Mobility-induced packet losses r Mobility patterns may be different m Controlled e.g. robots Offers opportunities for improving the network functions e.g. connectivity, coverage m Uncontrolled e.g. nomadic users Offers challenges to network design But also offers opportunities for improvement, e.g. –Users “carry” delay-tolerant data closer to destination –Delay Tolerant Network (Challenge Networks)

9 7: Wireless Ad Hoc Networks QoS r Providing QoS even in wired networks (e.g. the Internet) is a challenging problem r Wireless RF channels further complicate the problem m Unpredictability m Medium access: broadcast medium with hidden terminal problem r Possible solutions: m New MAC design m Cross-layer integration: allow different layers to adapt depending on available information at other layers

10 7: Wireless Ad Hoc Networks Scalability r Limited wireless transmission range r Whether the network is able to maintain an acceptable level of service even as the number of nodes is increased m How fast the network protocol control overhead increases as N increases r Possible solutions: m Introducing hierarchy m Utilizing location information m Limiting reactions to changes m Fixing things (e.g. paths) locally

11 7: Wireless Ad Hoc Networks Utilizing New Technologies r What are the gains that could be achieved by using newly available technologies such as m Smart directional (beamforming) antennas Increases the spatial reuse in cellular, but how about ad- hoc? Can several nodes together act as an antenna array? Practical issues? m Software Radio The ability to quickly switch the operating frequency may provide opportunities, but also challenging m GPS Location information may help

12 7: Wireless Ad Hoc Networks Security r Ease of snooping on wireless transmissions r From crypto point of view, lack of a trusted authority is one of the main challenges m How to generate/share keys reliably r Harder to track or even detect attackers in a wireless environment, given that: m Network relies on in-situ connections to other nodes which may be malicious r Malicious nodes may be especially harmful by injecting bogus control packets r DoS attacks that deplete a node’s battery

13 7: Wireless Ad Hoc Networks Lack of Reference r Lack of sufficient experimental data to confirm models m What does a multi-hop path really mean? m What is a link? r Simplistic models that do not capture the complexities, or complex models that do not lead to insights? r Are the protocols good enough, have they reached closed to the best possible? r Good balance between mathematical and experimental work

14 7: Wireless Ad Hoc Networks7-14 Multiple-Layer Problem r PHY m Adapt to rapid changes in link characteristics r MAC m Minimize collision, allow fair access, and semi-reliably transport under rapid change and hidden/exposed terminals r Network m Determine efficient transmission paths when links change often and bandwidth is at a premium r Transport m Handle delay and packet loss statistics that are very different than wired networks r Application m Handle frequent disconnection and reconnection as well as varying delay and packet loss characteristics

15 7: Wireless Ad Hoc Networks7-15 Several Major Issues r MAC protocols for ad hoc networks r Routing in ad hoc networks r Transport protocols for ad hoc networks

16 7: Wireless Ad Hoc Networks7-16 Design Goals for MAC Protocols r Allow fair access to the shared radio medium m Distributed protocol m Available bandwidth must be utilized efficiently m Control overhead should be minimized m Ensure fair bandwidth allocation to competing nodes m Reduce the effect of hidden/exposed terminals m Effectively manage the power consumption m Provide QoS support for real-time traffic m Protocol should be scalable

17 7: Wireless Ad Hoc Networks7-17 Overall Picture MAC Protocols for Ad Hoc Contention-based with reservation Contention-based with scheduling Other Protocols Sender initiatedReceiver initiated Single channelMultiple channel synchronousasynchronous DPS DWOP DLPS MMAC MCSMA PCM RBAR D-PRMA CATA HRMA SRMA/PA FPRP MACA/PR RTMAC BTMA DBTMA ICSMA MACAW FAMA RI-BTMA MACA-BI MARCH

18 7: Wireless Ad Hoc Networks7-18 Contention-based Protocols with Reservations r Use a bandwidth reservation technique m Contention occurs only at resource reservation phase m Node gets an exclusive access to the media once bandwidth is reserved r D-PRMA m Distributed packet reservation multiple access protocol r SRMA/PA m Soft reservation multiple access with priority assignment r RTMAC m Real-time medium access control protocol

19 7: Wireless Ad Hoc Networks7-19 Contention-based Protocols with Scheduling r Focus on packet scheduling at the nodes and transmission scheduling of the nodes r DPS m Distributed priority scheduling r DWOP m Distributed wireless ordering protocol r DLPS m Distributed laxity-based priority scheduling

20 7: Wireless Ad Hoc Networks7-20 Contention-based Protocols w/o Reservation/Scheduling r MACA m Multiple access collision avoidance protocol r MACAW m Media Access Protocol for Wireless LAN r BTMA m Busy tone multiple access protocol r MARCH m Media access with reduced handshake

21 7: Wireless Ad Hoc Networks7-21 MACA: Multiple Access Collision Avoidance r Proposed as an alternative to CSMA/CA r Handle hidden and exposed terminal issues using RTS-CTS r RTS and CTS packets carry the expected duration of the data transmission m A node near the sender that hearing RTS do not transmit for a time to receive CTS m A node near the receiver after hearing CTS differs its transmission m If the neighbor hears the RTS only, it is free to transmit once the waiting interval is over neighborsenderneighborreceiver RTS CTS Data

22 7: Wireless Ad Hoc Networks7-22 MACAW: Enhancement of MACA r Issue 1: potential flow starvation due to BEB m Both S1 and S2 have the high volume of traffic, S1 seizes the channel first m Packets transmitted by S2 get collided and it doubles CW m The probability that S2 seizes the channel decreasing r Solution in MACAW m Packet header contains the field set to the current back-off value of the transmitting node m Node receiving this packet copies this value to its back-off counter m If all the nodes can hear each other, eventually they will have the same back-off counter (fairness) S1 AP S2 × BEBBEB copy S1-AP S2-AP023.82

23 7: Wireless Ad Hoc Networks7-23 MACAW (Cont.) r Issue 2: backoff calculation adjusts too rapidly m After every successful transmission, return to the case where all stations have a minimal backoff counter, and then must repeat a period of contention to increase the backoffs r Solution in MACAW m Gentler adjustment Upon a collision, the backoff interval is increased by a multiplicative factor (1.5) F inc (x) = MIN[l.5x, CW max ] Upon success it is decreased by 1 F dec (x) = MAX[x-1, CW min ]

24 7: Wireless Ad Hoc Networks7-24 MACAW (Cont.) r Issue 3: Neighbor receivers problem m When node A sends an RTS to B, while node C is receiving from D, node B cannot reply with a CTS, since B knows that D is sending to C r When the transfer from C to D is complete, node B can send a Request-to-send-RTS (RRTS) to node A m Node A may then immediately send RTS to node B A B C D

25 7: Wireless Ad Hoc Networks7-25 MACAW (Cont.) r This approach, however, does not work in the scenario below m Node B may not receive the RTS from A at all, due to interference with transmission from C A B C D

26 7: Wireless Ad Hoc Networks7-26 BTMA: Busy Tone Multiple Access r One of the earliest solutions for hidden terminal problem r Multi-channel protocol m Control channel: used for busy tone transmission m Data channel: used for data transmission r Three variants: m BTMA (Busy Tone Multiple Access) m DBTMA (Dual Busy Tone Multiple Access) m RI-BTMA (Receiver-Initiated BTMA)

27 7: Wireless Ad Hoc Networks7-27 BTMA (Cont.) r Basic idea m Node senses the control channel to check whether the busy tone is active If not, turns on busy tone signal and starts data transmission If yes, waits for a random period of time and repeats m Any node that senses the carrier on the incoming data channel also transmits a busy tone r Pros and Cons m Simple with extremely low collision probability m Bandwidth utilization is low (blocked in two-hop neighbor) m Multiple channels

28 7: Wireless Ad Hoc Networks7-28 Several Major Issues r MAC protocols for ad hoc networks r Routing in ad hoc networks r Transport protocols for ad hoc networks

29 7: Wireless Ad Hoc Networks7-29 Why is Routing in MANET Different? r No specific nodes dedicated for control r Host mobility m Link failure/repair due to mobility may have different characteristics than those due to other causes r Rate of link failure/repair may be high when nodes move fast r Different node characteristics m E.g. power constraints, multiple access issues r New performance criteria may be used m Route stability despite mobility m Energy consumption

30 7: Wireless Ad Hoc Networks7-30 Unicast Routing Protocols r Many protocols have been proposed m Some have been invented specifically for MANET m Others are adapted from previously proposed protocols for wired networks r No single protocol works well in all environments m Some attempts made to develop adaptive protocols

31 7: Wireless Ad Hoc Networks7-31 MANET Protocol Zoo r Topology based routing m Proactive approach, e.g., DSDV. m Reactive approach, e.g., DSR, AODV, TORA. m Hybrid approach, e.g., Cluster, ZRP. r Position based routing m Location Services: DREAM, Quorum-based, GLS, Home zone etc. m Forwarding Strategy: Greedy, GPSR, RDF, Hierarchical, etc.

32 7: Wireless Ad Hoc Networks7-32 Routing Protocols r Proactive protocols m Determine routes independent of traffic pattern m Traditional link-state and distance-vector routing protocols are proactive r Reactive (on-demand) protocols m Discover/maintain routes only when needed m Source-initiated route discovery r Hybrid protocols

33 7: Wireless Ad Hoc Networks7-33 Trade-Off r Latency of route discovery m Proactive protocols may have lower latency since routes are maintained at all times m Reactive protocols may have higher latency because a route from X to Y will be found only when X attempts to send to Y r Overhead of route discovery/maintenance m Reactive protocols may have lower overhead since routes are determined only if needed m Proactive protocols can (but not necessarily) result in higher overhead due to continuous route updating

34 7: Wireless Ad Hoc Networks7-34 Tradeoff (Cont.) r Which approach achieves a better trade-off depends on the traffic and mobility patterns m Reactive protocols may yield lower routing overhead than proactive protocols when communication density is low m Reactive protocols tend to loose more packets (assuming that network layer drops packets if a route is not known) m Proactive protocols perform better with high mobility and dense communication graph

35 7: Wireless Ad Hoc Networks7-35 Single Path vs. Multipath r Single path m Use one path from source to destination m Similar to wired routes m Advantages: Simple to implement m Disadvantages: Source must find a new route to destination if old one fails r Multipath m Use more than one path from source to destination m Advantages: Load balancing can occur Higher tolerance to link failures m Disadvantages: Adds complexity to receiver and sender

36 7: Wireless Ad Hoc Networks7-36 Short Hops vs. Long Hops r Research to date suggests short-hop m Provides lower energy consumption Lower transmission power needed due to shorter distance between nodes m Provides higher link capacity Higher received signal strength due to shorter distance between nodes r Long-hop intuitively should have less total delay due to m Less total hops m Smaller total processing delay

37 7: Wireless Ad Hoc Networks7-37 Some Existing Wireless Routing Protocols r DSDV r WRP r CGSR r STAR r OLSR r FSR r HSR r GSR r DSR r AODV r ABR r SSA r FORP r PLBR r CEDAR r ZRP r ZHLS r RABR r LBR r COSR r PAR r LAR r OLSB

38 7: Wireless Ad Hoc Networks7-38 Dynamic Source Routing (DSR) r Reactive, source-based r When node S wants to send a packet to node D, but does not know a route to D, node S initiates a route discovery r Source node S floods Route Request (RREQ) r Each node appends own identifier when forwarding RREQ

39 7: Wireless Ad Hoc Networks7-39 Route Discovery in DSR B A S E F H J D C G I K Z Y Represents a node that has received RREQ for D from S M N L

40 7: Wireless Ad Hoc Networks7-40 Route Discovery in DSR B A S E F H J D C G I K Represents transmission of RREQ Z Y Broadcast transmission M N L [S] [X,Y] Represents list of identifiers appended to RREQ

41 7: Wireless Ad Hoc Networks7-41 Route Discovery in DSR B A S E F H J D C G I K Node H receives packet RREQ from two neighbors: potential for collision Z Y M N L [S,E] [S,C]

42 7: Wireless Ad Hoc Networks7-42 Route Discovery in DSR B A S E F H J D C G I K Node C receives RREQ from G and H, but does not forward it again, because node C has already forwarded RREQ once Z Y M N L [S,C,G] [S,E,F]

43 7: Wireless Ad Hoc Networks7-43 Route Discovery in DSR B A S E F H J D C G I K Z Y M Nodes J and K both broadcast RREQ to node D Since nodes J and K are hidden from each other, their transmissions may collide N L [S,C,G,K] [S,E,F,J]

44 7: Wireless Ad Hoc Networks7-44 Route Discovery in DSR B A S E F H J D C G I K Z Y Node D does not forward RREQ, because node D is the intended target of the route discovery M N L [S,E,F,J,M]

45 7: Wireless Ad Hoc Networks7-45 Route Discovery in DSR r Destination D on receiving the first RREQ, sends a Route Reply (RREP) r RREP is sent on a route obtained by reversing the route appended to received RREQ r RREP includes the route from S to D on which RREQ was received by node D

46 7: Wireless Ad Hoc Networks7-46 Route Reply in DSR B A S E F H J D C G I K Z Y M N L RREP [S,E,F,J,D] Represents RREP control message RREP [S,C,G,K,D]

47 7: Wireless Ad Hoc Networks7-47 Dynamic Source Routing (DSR) r Node S on receiving RREP, caches the route included in the RREP r When node S sends a data packet to D, the entire route is included in the packet header m Hence the name source routing r Intermediate nodes use the source route included in a packet to determine to whom a packet should be forwarded

48 7: Wireless Ad Hoc Networks7-48 DSR Optimization: Route Caching r Each node caches a new route it learns by any means r When node S learns that a route to node D is broken m Uses another route from its local cache, if such a route to D exists in its cache m Otherwise, node S initiates route discovery by sending a route request r Intermediate node X on receiving a Route Request for some node D can send a Route Reply m If node X knows a route to node D r Use of route cache m Can speed up route discovery m Can reduce propagation of route requests

49 7: Wireless Ad Hoc Networks7-49 DSR Pros and Cons r Advantages: m Less memory storage needed at each node since full routing table is not needed m Lower overhead needed because no periodic update message are necessary m Nodes do not need to continually inform neighbors they are still operational r Disadvantages: m Possible transmission latency due to reactive approach m Stale routes can occur if links change frequently m Message size increases as path length increases m Collisions between route requests propagated by neighboring nodes m Route Reply Storm due to nodes replying using their local cache

50 7: Wireless Ad Hoc Networks7-50 Several Major Issues r MAC protocols for ad hoc networks r Routing in ad hoc networks r Transport protocols for ad hoc networks

51 7: Wireless Ad Hoc Networks7-51 Transmission Control Protocol (TCP) r Reliable ordered delivery r Implements congestion avoidance and control r Reliability achieved by means of retransmissions if necessary r End-to-end semantics m Acknowledgements sent to TCP sender confirm delivery of data received by TCP receiver m Ack for data sent only after data has reached receiver

52 7: Wireless Ad Hoc Networks7-52 Challenges r Throughput unfairness m Unfairness at MAC layer m Transport layer should take this into account r Resource constraints m Power and bandwidth constraints r Separation of congestion control and reliability control r Completely decoupled transport layer m Wired network: transport protocol completely separated from underlying layer m Ad hoc network: interaction with network and MAC layer is expected for adaptability

53 7: Wireless Ad Hoc Networks7-53 Challenges (Cont.) r Misinterpretation of congestion m Traditional mechanism: packet loss, timeout m Ad hoc loss/delay due to High bit error rate due to varying link condition Packet collisions due to contention and hidden terminal Path breaks due to node mobility r Dynamically changing topology m Frequent path breaks m Partitioning and merging of networks m High delay in reestablishment of path

54 7: Wireless Ad Hoc Networks7-54 Performance of TCP r Several factors affect TCP performance in MANET m Wireless transmission errors m Multi-hop routes on shared wireless medium For instance, adjacent hops typically cannot transmit simultaneously m Route failures/changes due to mobility

55 7: Wireless Ad Hoc Networks7-55 Throughput over Multi-Hop Wireless Paths r Connections over multiple hops are at a disadvantage compared to shorter connections, because they have to contend for wireless access at each hop TCP Throughput using 2 Mbps MAC

56 7: Wireless Ad Hoc Networks7-56 Impact of Caching r Route caching has been suggested as a mechanism to reduce route discovery overhead r Each node may cache one or more routes to a given destination r When a route from S to D is detected as broken, node S may: m Use another cached route from local cache, or m Obtain a new route using cached route at another node

57 7: Wireless Ad Hoc Networks7-57 Why Performance Degrades With Caching r When a route is broken, route discovery returns a cached route from local cache or from a nearby node r After a time-out, TCP sender transmits a packet on the new route However, the cached route has also broken after it was cached r Another route discovery, and TCP time-out interval r Process repeats until a good route is found timeout due to route failure timeout, cached route is broken timeout, second cached route also broken

58 7: Wireless Ad Hoc Networks7-58 To Cache or Not to Cache r Caching can result in faster route “repair” r Faster does not necessarily mean correct r If incorrect repairs occur often enough, caching performs poorly r Need mechanisms for determining when cached routes are stale

59 7: Wireless Ad Hoc Networks7-59 Caching and TCP performance r Caching can reduce overhead of route discovery even if cache accuracy is not very high r But if cache accuracy is not high enough, gains in routing overhead may be offset by loss of TCP performance due to multiple time-outs

60 7: Wireless Ad Hoc Networks7-60 How to Improve Throughput (Bring Closer to Ideal) r Network feedback r Inform TCP of route failure by explicit message r Let TCP know when route is repaired m Probing m Explicit notification r Reduces repeated TCP timeouts and backoff

61 7: Wireless Ad Hoc Networks7-61 TCP with ELFN r Explicit Link Failure Notification m Not totally new, e.g., ECN bits in TCP m To provide the TCP sender with information about link and route failures, so that it can avoid responding to the failures as if congestion has occurred r How does it work? m When a TCP sender receives an ELFN: disables its retransmission timers and enters a “stand-by” mode m While on standby: A packet is sent at periodic intervals to probe the network to see if a route has been established m If an acknowledgment is received: leaves stand-by mode and restores the retransmission timers

62 7: Wireless Ad Hoc Networks7-62 Performance with Explicit Notification

63 7: Wireless Ad Hoc Networks7-63 Issues: Network Feedback r Network knows best (why packets are lost) Network feedback beneficial  Need to modify transport & network layer to receive/send feedback  Need mechanisms for information exchange between layers r [Holland99] discusses alternatives for providing feedback (when routes break and repair) m [Chandran98] also presents a feedback scheme

64 7: Wireless Ad Hoc Networks7-64 TCP Performance r Two factors result in degraded throughput in presence of mobility r Loss of throughput that occurs while waiting for TCP sender to timeout m This factor can be mitigated by using explicit notifications and better route caching mechanisms r Poor choice of congestion window and RTO values after a new route has been found m How to choose cwnd and RTO after a route change?

65 7: Wireless Ad Hoc Networks7-65 Issues: Window Size After Route Repair r Same as before route break: may be too optimistic r Same as startup: may be too conservative r Better be conservative than overly optimistic m Reset window to small value after route repair m Let TCP figure out the suitable window size m Impact low on paths with small delay-bw product

66 7: Wireless Ad Hoc Networks7-66 Issues: RTO After Route Repair r Same as before route break m If new route is long, this RTO may be too small, leading to timeouts r Same as TCP start-up (6 second) m May be too large m May result in slow response to next packet loss r Another plausible approach m RTO new = f(RTO old, route-length old, route-length new ) m E.g.: RTO new = RTO old * route-length new /route-length old m Not evaluated yet m Pitfall: RTT is not just a function of route length

67 7: Wireless Ad Hoc Networks7-67 Summary r Still many remained topics related to wireless ad hoc networks r More research opportunities in m Wireless mesh networks With fixed infrastructure as wireless infrastructure Multi-radio multi-channel architecture m Wireless sensor networks Energy consumption is one of the key challenges Application specific demands, including localization, coverage, event detection/collection, etc.


Download ppt "7: Wireless Ad Hoc Networks7-1 Chapter 7 Wireless Ad Hoc Networks."

Similar presentations


Ads by Google