sensor and energy-efficient networking
DESCRIPTION
Sensor and energy-efficient networking. CSE 525: Advanced Networking Computer Science and Engineering Department Winter 2004. Energy efficient MAC. An energy-efficient MAC protocol for wireless sensor networks by W. Ye, J. Heidemann, and D. Estrin - PowerPoint PPT PresentationTRANSCRIPT
Oregon Graduate Institute 1
Sensor and energy-efficient networking
CSE 525: Advanced NetworkingComputer Science and Engineering Department
Winter 2004
Oregon Graduate Institute 2
Energy efficient MAC
An energy-efficient MAC protocol for wireless sensor networks by W. Ye, J. Heidemann, and D. Estrin
An Energy Efficient MAC Protocol for Wireless LANs by E. Jung and N. Vaidya
Energy efficient communications in ad hoc networks using directional antennas by A. Spyropoulos and C. Raghavendra
Oregon Graduate Institute 3
Motivation
Challenged Networks are normally battery operated, hence power limited
Other goals: Self-configuration good scalability collision avoidance
Fairness and latency are NOT important
Oregon Graduate Institute 4
Conventional MAC Layer An access mechanism for nodes that
ensure that NO two nodes have access to the communication channel concurrently
If air is clear, send; for receiving, wait
How do we know when to receive? Listen always!
Usually designed to allow for maximum throughput
Hence not energy-efficient
Oregon Graduate Institute 5
Sources of energy inefficiency
Collision Overhearing Control packet overhead Idle listening
Oregon Graduate Institute 6
#1: S(sensor)-MAC It is contention based Tries to reduce wastage of energy from
all four sources of energy inefficiency Collision – by using RTS and CTS Overhearing – by switching the radio off
when the transmission is not meant for that node
Control overhead – by Message Passing Idle listening – by Periodic Sleep
Oregon Graduate Institute 7
There is no FREE dinner!
In exchange there is some reduction in both per-hop fairness and latency
But this does not necessarily result in lower end-to-end application fairness and latency – Which acceptable in Challenged Network
Oregon Graduate Institute 8
Periodic Listen and Sleep
Each node go into periodic sleep mode during which it switches the radio off and sets a timer to awake later
To reduce control overhead, neighboring nodes are synchronized (i.e. Listen and sleep together)
Oregon Graduate Institute 9
Periodic Listen and Sleep Broadcasting schedule to all its
immediate neighbor neighbor can have different schedules.
If multiple neighbor want to talk to a node, use 802.11 (RTS/CTS) like contention scheme
After they start data transmission, they do not follow their sleep schedules until they finish transmission.
Oregon Graduate Institute 10
Choosing and Maintaining Schedules
Each node maintains a schedule table that stores schedules of all its known neighbors.
To establish the initial schedule (at the startup) following steps are followed: A node first listens for a certain amount of
time. If it does not hear a schedule from another
node, it randomly chooses a schedule and broadcast its schedule immediately.
This node is called a SYNCHRONIZER.
Oregon Graduate Institute 11
If a node receives a schedule from a neighbor before choosing its own schedule, it just follows this neighbor’s schedule.
This node is called a FOLLOWER and it waits for a random delay and broadcasts its schedule.
Choosing and Maintaining Schedules
Oregon Graduate Institute 12
Border Nodes
If a node receives a different schedule after it selects and broadcasts its own schedule, it adopts both schedules
Border nodes consume more energy
Oregon Graduate Institute 13
Maintaining Synchronization
Timer synchronization among neighbors are needed to prevent the clock drift.
Synchronizer needs to periodically send SYNC to its followers.
If a follower has a neighbor that has a different schedule with it, it also needs update that neighbor.
Oregon Graduate Institute 14
Maintaining Synchronization
Time of next sleep is relative to the moment that the sender finishes transmitting the SYNC packet
Receivers will adjust their timer counters immediately after they receive the SYNC packet
Listen interval is divided into two parts: one for receiving SYNC and other for receiving RTS
Oregon Graduate Institute 15
Timing Relationship of Possible Situations
Oregon Graduate Institute 16
Collision/Overhearing Avoidance Collision Avoidance: 802.11
RTS/CTS NAV: virtual carrier sense Physical carrier sense
Overhearing Avoidance Go to sleep if overhear RTS/CTS packet
NAV Avoid overhearing “long” data packet
All immediate neighbors of both sender and receiver go to sleep
Oregon Graduate Institute 17
Message Passing Short packets are used in wireless networks
More robust Overhead of control packets (RTS/CTS) In-network processing requires a complete msg
Divide the long message into small fragments and transmit them in a burst. RTS/CTS/Data1/Ack1/Data2/Ack2/…/DataN/AckN If a packet is lost, extend the duration and
immediately retransmit it Data and Acknowledge include a field of duration
Oregon Graduate Institute 18
Energy Savings vs. Increased Latency
Delay includes: Carrier sense Back off Transmission Propagation Processing Queuing Sleeping
Oregon Graduate Institute 19
Conclusions and Future work
S-MAC has good energy conserving properties comparing to IEEE 802.11
Future work Analytical study on the energy
consumption and latency Analyze the effect of topology
changes
Oregon Graduate Institute 20
#2: EEM for WLAN Optimize PSM in the DCF in IEEE 802.11
Standard Dynamic PSM (DPSM)
What is PSM? PSM for DCF, divide time into intervals called
beacon intervals, each node in power save mode periodically wakes up at the beginning of beacon interval for a duration called ATIM [Ad-hoc Traffic Indication Message] window to exchange control information.
Oregon Graduate Institute 21
What’s wrong?
Fixed ATIM window does not perform good in all situation Adaptive mechanism to dynamically
adjust ATIM window Synchronization of beacon interval
of initially partitioned network Not addressed; it assumes, there is a
way to synchronize
Oregon Graduate Institute 22
Oregon Graduate Institute 23
Dynamic - PSM Each node chose its own ATIM window
size based on network traffic condition Allow to increase and decrease the
ATIM window dynamically; ATIMmin is defined as minimum level
Move into doze state, after completing packet transmission, if remaining time is not “too-small” – Save Energy
Oregon Graduate Institute 24
Dynamic - PSM Piggyback own window size on all
transmitted packets Packet marking use to adjust ATIM window
Oregon Graduate Institute 25
Dynamic - PSM Rules to increase ATIM
Pending packet can’t be announce in current window time
Based on piggyback information Receiving an ATIM frame after ATIM window Received a marked packet
Rules to decrease ATIM If node successfully announce ATIM frame
and none of the above rule satisfied
Oregon Graduate Institute 26
Result
Simulation result shows that DPSM improves energy consumption without degrading performance Only when energy gain from doze
state is more then energy loss [overhead of beacon, ATIM and ATIM-ACK frame]
Energy save in doze mode is 96% compare to idle mode.
Oregon Graduate Institute 27
#3: Directional Antenna Energy efficient routing and scheduling
algorithm in ad hoc network where each node has single directional antenna.
Using topology consisting of all the possible link in the network, find shortest cost path to be energy efficient.
Calculate the amount of traffic that has to go over each link and find the maximum amount of time each link can be up.
Schedule node’s transmissions, trying to minimize the total time it takes for all possible Tx-Rx pairs to communicate with each other.
Oregon Graduate Institute 28
Conclusion and future work Benefits of using directional antennas in
ad hoc networks. Energy efficient algorithm for routing. Scheduling
45% improvement in network life time that is achieved by using energy-aware routing.
Future work Multicasting and broadcasting in ad hoc
networks with directional antennas Scheduling algorithm in this context
Oregon Graduate Institute 29
Related work TDMA Based
Naturally have a duty cycle It is not easy to change the slot assignment
dynamically, hence scalability is not as good as contention based
Requires nodes to form real communication clusters and managing inter-cluster communication is difficult
Out-of band solutions [PAMAS]: Requires extra band for signaling
Oregon Graduate Institute 30
Questions?