Lecture12-Nov09-05 - M edi um Access Contr ol i n Ad hoc a...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: M edi um Access Contr ol i n Ad hoc a nd Sensor N etwor ks M ul ti pl e Access Contr ol (M AC) Pr otocol s MAC allows multiple users to share a common channel. Conflict-free protocols ensure successful transmission. Channel can be allocated to users statically or dynamically. Only static conflict-free protocols are used in cellular mobile communications - Frequency Division Multiple Access (FDMA): provides a fraction of the frequency range to each user for all the time - Time Division Multiple Access (TDMA) : The entire frequency band is allocated to a single user for a fraction of time - Code Division Multiple Access (CDMA) : provides every user a portion of bandwidth for a fraction of time Contention based protocols must prescribe ways to resolve conflicts - Static Conflict Resolution: Carrier Sense Multiple Access (CSMA) - Dynamic Conflict Resolution: the Ethernet, which keeps track of various system parameters, ordering the users accordingly Fr equency D i vi si on M ul ti pl e Access (FD M A) Channels are assigned to the user for the duration of a call. No other user can access the channel during that time. When call terminates, the same channel can be re-assigned to another user FDMA is used in nearly all first generation mobile communication systems, like AMPS (30 KHz channels) Number of channels required to support a user population depends on the average number of calls generated, average duration of a call and the required quality of service (e.g. percentage of blocked calls) Channel 1 Bandwidth Channel 2 Channel 3 Channel 4 Time T i me D i vi si on M ul ti pl e Access (T D M A) The whole channel is assigned to each user, but the users are multiplexed in time during communication. Each communicating user is assigned a particular time slot, during which it communicates using the entire frequency spectrum The data rate of the channel is the sum of the data rates of all the multiplexed transmissions There is always channel interference between transmission in two adjacent slots because transmissions tend to overlap in time. This interference limits the number of users that can share the channel Channel 3 Channel 3 Channel 3 Channel 2 Channel 2 Channel 1 Channel 1 Channel 2 Channel 2 Channel 4 Channel 4 Channel 1 Bandwidth Time Code D i vi si on M ul ti pl e Access (CD M A) CDMA, a type of a spread-spectrum technique, allows multiple users to share the same channel by multiplexing their transmissions in code space. Different signals from different users are encoded by different codes (keys) and coexist both in time and frequency domains A code is represented by a wideband pseudo noise (PN) signal When decoding a transmitted signal at the receiver, because of low crosscorrelation of different codes, other transmissions appear as noise. This property enables the multiplexing of a number of transmissions on the same channel with minimal interference The maximum allowable interference (from other transmissions) limits the number of simultaneous transmissions on the same channel Bandwidth All channels share bandwidth Time Code D i vi si on M ul ti pl e Access (CD M A) Spreading of the signal bandwidth can be performed using Direct Sequence (DS): The narrow band signal representing digital data is multiplied by a wideband pseudo noise (PN) signal representing the code. Multiplication in the time domain translates to convolution in the spectral domain. Thus the resulting signal is wideband Frequency Hopping (FH): Carrier frequency rapidly hops among a large set of possible frequencies according to some pseudo random sequence (the code). The set of frequencies spans a large bandwidth. Thus the bandwidth of the transmitted signal appears as largely spread An Ener gy-Effi ci ent M AC Pr otocol for Wi r el ess Sensor N etwor ks (S-M AC) [Ye+ 2002] S- MAC protocol designed specifically for sensor networks to reduce energy consumption while achieving good scalability and collision avoidance by utilizing a combined scheduling and contention scheme The major sources of energy waste are: 1. 2. 3. 4. collision overhearing control packet overhead idle listening S-MAC reduce the waste of energy from all the sources mentioned in exchange of some reduction in both per-hop fairness and latency (S-M AC) S- MAC protocol consist of three major components: 1. 2. 3. periodic listen and sleep collision and overhearing avoidance Message passing Contributions of S-MAC are: The scheme of periodic listen and sleep helps in reducing energy consumption by avoiding idle listening. The use of synchronization to form virtual clusters of nodes on the same sleep schedule In-channel signaling puts each node to sleep when its neighbor is transmitting to another node (solves the overhearing problem and does not require additional channel) Message passing technique to reduce application-perceived latency and control overhead (per-node fragment level fairness is reduced) Evaluating an implementation of S-MAC over sensor-net specific hardware A Power Contr ol M AC (PCM ) Pr otocol for Ad hoc N etwor ks [J ung+ 2002] A power control MAC protocol allows nodes to vary transmit power level on a per-packet basis Earlier work has used different power levels for RTS-CTS and DATA-ACK, specifically, maximum transmit power is used for RTS-CTS and minimum required transmit power is used for DATA-ACK transmissions These protocols may increase collisions, degrade network throughput and result in higher energy consumption than when using IEEE 802.11 without power control Power saving mechanisms allow nodes to enter a doze state by powering off its wireless network interface whenever possible Power control schemes vary transmit power to reduce energy consumption Power Contr ol M AC (PCM ) IEEE 802.11 MAC Protocol Specifies two MAC protocols: Point Coordination Function (PCF) centralized Distributed Coordination Function (DCF) distributed Transmission range: When a node is in transmission range of a sender node, it can receive and correctly decode packets from sender node. Carrier Sensing Range: Nodes in carrier sensing range can sense the sender's transmission. It is generally larger than transmission range. Both carrier sensing range and transmission range Depends on the transmit power level. Power Contr ol M AC (PCM ) IEEE 802.11 MAC Protocol Carrier Sensing Zone: Nodes can sense the signal, but cannot decode it correctly. The carrier sensing zone does not include transmission range [Figure adapted from Jung+ 2002] Power Contr ol M AC (PCM ) IEEE 802.11 MAC Protocol DCF in IEEE 802.11 is based on CSMA/CS (Carrier Sense Multiple Access with Collision Avoidance) Each node in IEEE 802.11 maintains a NAV (Network Allocation Vector) that indicates the remaining time of the on-going transmission sessions Carrier sensing is performed using physical carrier sensing (by air interface) and virtual carrier sensing (uses the duration of the packet transmission that is included in the header of RTS, CTS and DATA frames) Using the duration information in RTS, CTS and DATA packets, nodes update their NAVs whenever they receive a packet The channel is considered busy if either physical or virtual carrier sensing indicates that channel is busy Figure 2 shows how nodes in transmission range and the carrier sensing zone adjust their NAVs during RTS-CTS-DATA-ACK transmission Power Contr ol M AC (PCM ) IEEE 802.11 MAC Protocol [Figure adapted from Jung+ 2002] Power Contr ol M AC (PCM ) IEEE 802.11 MAC Protocol IFS is the time interval between frames and IEEE 802.11 defines four IFSs which provide priority levels for accessing the channel SIFS (short interframe space) PIFS (Point Coordination Function interframe space) DIFS (Distributed Coordination Function interframe space) EIFS (extended interframe space) SIFS is the shortest and is used after RTS, CTS, and DATA frames to give the highest priority to CTS, DATA and ACK respectively In DCF, when the channel is idle, a node waits for DIFS duration before transmitting Nodes in the transmission range correctly set their NAVs when receiving RTS/CTS Since nodes in carrier sensing zone cannot decode the packet, they do not know the duration of the packet transmission. So, they set their NAVs for the EIFS duration to avoid collision with the ACK reception at the source node Power Contr ol M AC (PCM ) IEEE 802.11 MAC Protocol The intuition behind EIFS is to provide enough time for a source node to receive the ACK frame, meaning that duration of EIFS is longer than that of ACK transmission In PCM, nodes in the carrier sensing zone use EIFS whenever they can sense the signal but cannot decode it IEEE 802.11 does not completely prevent collisions due to the hidden terminal problem (nodes in the receiver's carrier sensing zone, but not in the sender's carrier sensing zone or transmission range, can cause a collision with the reception of a DATA packet at the receiver In Figure 3, suppose node C transmits packet to node D When C and D transmit an RTS and CTS respectively, A and F set their NAVs for EIFS duration During C's data transmission, A defers its transmission due to sensing C's transmission. However, since node F does not sense any signal during C's transmission, it considers channel to be idle (F is in D's carrier sensing zone, but not in D's) Power Contr ol M AC (PCM ) IEEE 802.11 MAC Protocol C's carrier sensing range D's carrier sensing range [Figure adapted from Jung+ 2002] Power Contr ol M AC (PCM ) IEEE 802.11 MAC Protocol When F starts a new transmission, it can cause a collision with the reception of DATA at D Since F is outside of D's transmission range, D may be outside of F's transmission range; however, since F is in D's carrier sensing zone, F can provide interference at node D to cause collision with DATA being received at D Power Contr ol M AC (PCM ) BASIC Power Control Protocol Power control can reduce energy consumption Power control may bring different transmit power levels at different hosts, creating an asymmetric scenarios where a node A can reach node B, but node B cannot reach node A and collisions may also increase a result In Figure 4, suppose nodes A and B use lower power level than nodes C and D When A is transmitting to B, C and D may not sense the transmission When C and D transmit to each other using higher power, their transmission may collide with the on-going transmission from A to B [Figure adapted from Jung+ 2002] Power Contr ol M AC (PCM ) BASIC Power Control Protocol As a solution to this problem, RTS-CTS are transmitted at the highest possible power level but DATA and ACK at the minimum power level necessary to communicate In Figure 5, nodes A and B send RTS and CTS respectively with highest power level such that node C receives the CTS and defers its transmission By using a lower power level for DATA and ACK packets, nodes can save energy [Figure adapted from Jung+ 2002] Power Contr ol M AC (PCM ) BASIC Power Control Protocol In the BASIC scheme, RTS-CTS handshake is used to decide the transmission power for subsequent DATA and ACK packets which can be achieved in two different ways Suppose node A wants to send a packet to node B. Node A transmit RTS at power level pmax (maximum possible). When B receives the RTS from A with signal level pr, B calculates the minimum necessary transmission power level, pdesired. For the DATA packet based on received power level, pr, transmitted power level, pmax, and noise level at the receiver B. Node B specifies pdesired in its CTS to node A. After receiving CTS, node A sends DATA using power level pdesired. When a destination node receives an RTS, it responds by sending a CTS (at power level pmax). When source node receives CTS, it calculates pdesired based on received power level, pr, and transmitted power level (pmax) as Pdesired = (pmax / pr) x Rxthresh x c where Rxthresh is minimum necessary received signal strength and c is constant Power Contr ol M AC (PCM ) BASIC Power Control Protocol The second alternative makes two assumptions: Signal attenuation between source and destination nodes is assumed to be the same in both directions Noise level at the receiver is assumed to be below some predefined threshold Deficiency of the BASIC Protocol In Figure 6, suppose node D wants to transmit to node E When nodes D and E transmits RTS and CTS respectively, B and C receives RTS and F and G receives CTS, therefore, these nodes defer their transmissions Since node A is in carrier sensing zone of node D, it sets its NAV for EIFS duration Similarly node H sets its NAV for EIFS duration when it senses transmission from E When source and destination decide to reduce the transmit power for DATA-ACK, not only transmission range for DATA-ACK but also carrier sensing zone is also smaller than RTS-CTS Power Contr ol M AC (PCM ) Deficiency of the BASIC Protocol Thus, only C and F correctly receives DATA and ACK packets Since nodes A and H cannot sense the transmissions, they consider channel is idle and start transmitting at high power level which will cause collision with the ACK packet at D and DATA packet at E This results in throughput degradation and higher energy consumption (due to retransmissions) [Figure adapted from Jung+ 2002] Power Contr ol M AC (PCM ) Proposed Power Control MAC Protocol Proposed Power Control MAC (PCM) is similar to BASIC scheme such that it uses power level, pmax, for RTS-CTS and the minimum necessary transmit power for DATA-ACK transmissions Procedure of PCM is as follows: 1. 2. 3. Source and destination nodes transmit the RTS and CTS using pmax. Nodes in the carrier sensing zone set their NAVs for EIFS duration The source may transmit DATA using a lower power level Source transmits DATA at level of pmax, periodically, for enough time so that nodes in the carrier sensing zone can sense it and this would avoid collision with the ACK packets 4. The destination node transmits an ACK using the minimum required power to reach the source node Figure 7 presents how the transmit power level changes during the sequence of RTS-CTS-DATA-ACK transmission Power Contr ol M AC (PCM ) Proposed Power Control MAC Protocol The difference between PCM and BASIC scheme is that PCM periodically increases the transmit power to pmax during the DATA packet transmission. Nodes that can interfere with the reception of ACK at the sender will periodically sense the channel is busy and defer their own transmission. Since nodes reside in the carrier sensing zone defer for EIFS duration, the transmit power for DATA is increased once every EIFS duration PCM solves the problem posed with BASIC scheme and can achieve throughput comparable to 802.11 by using less energy PCM, like 802.11, does not prevent collisions completely [Figure adapted from Jung+ 2002] A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Why STUDY MAC protocols in sensor networks? Application behavior in sensor networks leads to very different traffic characteristics from that found in conventional computer networks Highly constrained resources and functionality Small packet size Deep multi-hop dynamic topologies The network tends to operate as a collective structure, rather than supporting many independent point-to-point flows Traffic tends to be variable and highly correlated Little or no activity/traffic for longer periods and intense traffic over shorter periods A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Major factors to be considered in the design of MAC: Communication efficiency in terms of energy consumed per each packet o o Communication by radio channel consumes the highest energy Transmit , receive and idle consume roughly the same amount of energy Fairness of the bandwidth allocated to each node for end to end data delivery to sink o o o Each node acts as a router as well as data originator resulting in two kinds of traffic The traffics compete for the same upstream bandwidth Hidden nodes Contention at the upstream node may not be detected and results in significant loss rate Efficient channel utilization A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Major factors to be considered in the design of MAC: The routing distance and degree of intermediate competition varies widely across the network The cost of dropping a packet varies with place and the packet Listening mechanism: o o o o o o Listening is effective when there are no hidden nodes It comes at an expense of energy cost as the radio must be on to listen Many protocols such as IEEE 802.11 require sensing the channel even during backoff Shorten the length of carrier sensing and power off the node during backoff Highly synchronized nature of the traffic causes no packet transfer at all in the absence of collision detection hardware Introduce random delay for transmission to unsynchronized the nodes Contribution of this paper are as follows: A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Backoff Mechanism: o o o Used to reduce the contention among the nodes In the sensor networks, traffic is a superposition of different periodic streams Apply back off as a phase shift to the periodicity of the application so that the synchronization among periodic streams of traffic can be broken Contention-based Mechanism o o o o Explicit control packets like RTS and CTS are used to avoid contention ACKS indicate lack of collision Use of lot of control packets reduces bandwidth efficiency ACKS can be eliminated by hearing the packet transmission from its parent to its upstream which serves as an ACK for the downstream node A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Rate Control Mechanism o o The competition between originating traffic and route-thru traffic has a direct impact in achieving the fairness goal. MAC should control the rate of originating data of a node in order to allow route-thru traffic to access the channel and reach the base station and some kind of progressive signaling for route-thru traffic such the rate is controlled at the origin. A passive implicit mechanism is used to control the rate of transmission of both traffics o A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Multi-hop Hidden Node problem: o It avoid the hidden node problem by constantly tuning the transmission rate and performing phase changes so that the aggregate traffic will not repeatedly collide with each other. A child can reduce a potential hidden node problem with its grand parent by not sending packets for t+ x+ packet time at the end of packet transmission t by its parent o A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Advantages: The amount of computation for this scheme is small and within networked sensor's computation capability The scheme is totally computational which is much cheaper in energy cost than on the radio The control packet overhead is reduced A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Disadvantages: The MAC protocol developed here takes into consideration the periodicity of the originating traffic which doesn't help for non periodic traffic A T r a nsmi ssi on Contr ol Scheme for M edi a Access i n Sensor N etwor ks [Woo+, 2003] Suggestions/Improvements/Future Work: An Ada pti ve Ener gy-Effi ci ent M AC Pr otocol for Wi r el ess Sensor N etwor ks [Va n da m+, 2003] T-MAC is a contention based Medium Access Control Protocol Energy consumption is reduced by introducing an active/sleep duty cycle Handles the load variations in time and location by introducing an adaptive duty cycle It reduces the amount of energy wasted on idle listening by dynamically ending the active part of it In T-MAC, nodes communicate using RTS, CTS, Data and ACK pkts which provides collision avoidance and reliable transmission When a node senses the medium idle for TA amount of time it immediately switches to sleep TA determines the minimal amount of idle listening time per frame The incoming messages between two active states are buffered An Ada pti ve Ener gy-Effi ci ent M AC Pr otocol for Wi r el ess Sensor N etwor ks [Va n da m+, 2003] The buffer capacity determines an upper bound on the maximum frame time Frame synchronization in T-MAC follows the scheme of virtual clustering as in S-MAC The RTS transmission in T-MAC starts by waiting and listening for a random time within a fixed contention interval at the beginning of the each active state The TA time is obtained using TA > C + R + T T-MAC suffers from early sleeping problem Its overcome by sending Future request to send or taking priority on full buffers An Ada pti ve Ener gy-Effi ci ent M AC Pr otocol for Wi r el ess Sensor N etwor ks [Va n da m+, 2003] Advantages: The T-MAC protocol is designed particularly for wireless sensor networks and hence energy consumption constraints are taken into account The T-MAC protocol tries to reduce idle listening by transmitting all messages in bursts of variable lengths and sleeping between burst T-MAC facilitates collision avoidance and overhearing -- nodes transmit their data in a single burst and thus do not require additional RTS/CTS control packets. By stressing on RTS retries, T-MAC gives the receiving nodes enough chance to listen and reply before it actually goes to sleep -- this increases the throughput in the long run An Ada pti ve Ener gy-Effi ci ent M AC Pr otocol for Wi r el ess Sensor N etwor ks [Va n da m+, 2003] Disadvantages: The authors do not outline how a sender node would sense a FRTS packet and enable it to send a DS packet Also sending a DS packet increases the overhead. The network topology in the simulation considers that the locations of the nodes are known T-MAC has been observed to have a high message loss phenomenon T-MAC suffers from early sleeping problem for event based local unicast An Ada pti ve Ener gy-Effi ci ent M AC Pr otocol for Wi r el ess Sensor N etwor ks [Va n da m+, 2003] Suggestions/Improvements/Future Work: If a buffer is full there would be a lot of dropped packets decreasing the throughput. A method to overcome this drawback is that we could have the node with its buffer 75% full broadcast a special packet Buffer Full Packet MAC Virtual Clustering technique needs to be further investigated An adaptive election algorithm can be incorporated where the schedule and neighborhood information is used to select the transmitter and receivers for the current time slot, hence avoiding collision and increasing energy conservation Refer ences [Jung+ 2002] E.-S. Jung and N.H. Vaidya, A Power Control MAC Protocol for Ad hoc Networks, Proceedings of ACM MOBICOM 2002, Atlanta, Georgia, September 23-28, 2002. [Ye+ 2002] W. Yei, J. Heidemann and D. Estrin, Energy-Efficient MAC Protocol for Wireless Sensor Networks, Proceedings of the Twenty First International Annual Joint Conference of the IEEE Computer and Communications Societies (INFOCOM 2002), New York, NY, USA, June 23-27 2002. [Woo+ 2003] A. Woo and D. Culler, A Transmission Control Scheme for Media Access in Sensor Networks, Proceedings of the ACM/IEEE International Conference on Mobile Computing and Networking, Rome, Italy, July 2001, pp. 221-235. [Van Dam+ 2003] T. V. Dam and K. Langendoen, An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks, ACM SenSys, Los Angeles, CA, November, 2003. ...
View Full Document

This note was uploaded on 08/25/2011 for the course EEL 6788 taught by Professor Boloni,l during the Spring '08 term at University of Central Florida.

Ask a homework question - tutors are online