1
An Adaptive Energy-Efficient MAC Protocol for Wireless Sensor Networks
Tijs van Dam, Koen LangendoenIn ACM SenSys 2003.
8/1/2005Hong-Shi Wang
2
Contents
IntroductionRelated workDrawbacks of S-MACT-MACExperimentsConclusions and Future Work
3
Introductions
Traditional MAC Protocols– Design to maximize packet throughput, minimize latency and
provide fairness
Protocol design for wireless sensor networks– focuses on minimizing energy consumption
4
Related Work
TDMA-based protocol– Have advantage of energy conservation compared to
contention protocols, because there is no contention-introduced overhead and collisions
– But requires the nodes to form real communication clusters like LEACH Managing inter-cluster communication and interference is not an
easy task.
Contention-based protocol– simplicity– Energy consumption using this MAC is very high when nodes
are in idle mode
5
Drawbacks of S-MAC
Active (Listen) interval – If message rate is less – energy is still wasted in idle-listening
S-MAC’ fixed duty cycle is NOT OPTIMAL
6
T-MAC : Preliminaries (1)
Basic idea– To utilize an active and a sleep cycle, similar to S-MAC– To introduce an adaptive duty cycle by dynamically ending
the active part– An active period ends when no activation event has occurred
for a time TA Activation event
– The reception of any data on the radio (RTS, CTS, DATA, ACK)– The sensing of communication on the radio (overhearing)
– Difference in the duty cycle S-MAC - fixed duty cycle T-MAC – Dynamic duty cycle
7
T-MAC : Preliminaries (2)
With normal MAC protocols, messages are spread out over the whole time frame
With S-MAC, active time is fixed With T-MAC, the active time is dynamically adjusted (i.e., be
shorten) by timing out on hearing nothing during some time period (TA)– TA determines the minimal amount of idle listening per frame
Active Active Active
Sleep Sleep
S-MAC
Active Active Active
Sleep Sleep
TA TATAT-MAC
8
T-MAC : RTS Operation (1)
Contention Interval
In contention-based protocols, like IEEE 802.11, a back-off scheme is used: – Contention interval increases when traffic is higher.
In the T-MAC protocol, waiting and listening for a random time within a fixed contention interval – Tuned for maximum load.
9
T-MAC : RTS Operation (2)
RTS Retries
No CTS reply for RTS?– Collision– The receiving node is prohibited from replying due to an
overhead RTS or CTS– Receiving node is asleep
Solutions:– Retransmit RTS if no answer– If there is still no reply after two retries, it should give up and
go to sleep
10
T-MAC : Choosing TA
Determining TA
The interval TA must be long enough to receive at least the start of the CTS packet
TA > C+R+T– C –contention interval length; R–RTS packet length;– T –turn-around time, time between RTS end & CTS start– Larger TA increases the energy used– In experiments, used TA = 1.5 x (C + R + T)
A
B
C
contend
contend
RTS CTS DATA ACK
TA
11
T-MAC : Overhearing Avoidance
~= S-MAC But implemented as an option in T-MAC Node goes to sleep after overhearing RTS/CTS of
other nodes communication– miss other RTS/CTS while sleeping– throughput decreases
Although overhearing avoidance saves energy, it must not be used when maximum throughput is required
12
T-MAC: Asymmetric Communication (1)Early-Sleeping Problem – unidirectional (A to D) If node C looses contention because it overhears a CTS packet
from B to A, C must remain silent. Since D does not know of the communication between A and B,
its active time will end, and node D will go to sleep. Only at the start of the next frame will node C have a new
chance to send to node D Early-Sleeping Problem
– Node goes to sleep when a neighbor still has messages for it
A
B
C
D
contend
contend
RTS CTS DATA ACK
RTS?
active sleep
TA
13
T-MAC: Asymmetric Communication (2)Future request-to-send (FRTS) Let others know that we still have a message for it, but cannot
access the medium; – C sends FRTS to future target of an RTS packet
– FRTS has duration field
FRTS might affect data; so, DATA postponed until FRTS is over; To prevent others from taking medium, A send DS packet;
A
B
C
D
contend
contend
RTS CTS DATA ACK
RTS
active
TA
DS
FRTS
active
TA > C+R+T+CTS_length
14
T-MAC: Asymmetric Communication (3)Taking priority on full buffers When a node’s transmit/routing buffers are almost full, it may
prefer sending than receiving Receive RTS, send its own RTS to others instead of CTS limited form of flow control
A
B
C
D
contend
contend
RTS
CTS DATA ACK
active
RTS
contend
TA
15
Experiments
S-MAC Vs. T-MAC
16
Simulation setup and parameters
Simulator: OMNeT++ Built a network of 100 nodes in a 10 by 10 grid Energy consumption
–
S-MAC protocol– A frame length of one second, and with several lengths of the active
time, varying from 75 ms to 915 ms.
T-MAC protocol– Always used a frame length of 610ms and an interval TA with a len
gth of 15 ms
– Can optionally be deployed with overhearing avoidance, full-buffer priority, and FRTS
20 A while sleeping, 4mA while receiving and 10 mA while transmitting
17
Homogeneous local unicast
Nodes send packets to one of their neighbors at random T-MAC: Used overhearing avoidance, but no FRTS or full-buffer
priority mechanisms
18
Nodes-to-sink communication
Nodes send messages to a single sink node – Shortest path routing, no data aggregation
T-MAC: Used overhearing avoidance, FRTS or full-buffer priority mechanisms
19
Early-sleeping Problem Nodes send messages to a single sink node
– Shortest path routing, no data aggregation
T-MAC: FRTS Vs. PriorityVs. FRTS + PriorityVs. No measures
20
Event-Based Local Unicast When no events happen, nodes exchange local messages of 10 bytes
with each other every 20 seconds. Also report to a sink node every 100 seconds.
When an event happens, nodes start sending local unicast messages of 30 bytes. Then also send messages of 50 bytes to the sink.
21
Conclusions And Future Work
Conclusions T-MAC dynamically adapts a listen/sleep duty cycle T-MAC Protocol
– During a high load, nodes communicate without sleeping– During a very low load, nodes will use their radios for as little
as 2.5% of the time, saving as much as 96% of the energy compared to a traditional
Future Work Throughput and Early-Sleeping Problem
22
The End
Thanks for your listening !