Wireless Sensor Networks
Link Layer Protocols
Mario Čagalj
mario.cagalj@fesb.hr
FESB
University of Split
2013.
Based on “Protocols and Architectures for Wireless Sensor Networks”, Holger Karl, 2005.
1
Data Link Layer services (source: Wiki)
o Encapsulation of network layer data packets into frames
o Frame synchronization
o Logical link control (LLC) sublayer:
> Error control (automatic repeat request,ARQ), in addition to ARQ
provided by some Transport layer protocols, to forward error
correction (FEC) techniques provided on the Physical Layer, and to
error-detection and packet canceling provided at all layers, including
the network layer. Data link layer error control (i.e. retransmission of
erroneous packets) is provided in wireless networks, error detection
and canceling of erroneous packets are provided.
> Flow control, in addition to one provided on the Transport layer. Data
link layer error control is not used in LAN protocols such as
Ethernet, but in modems and wireless networks.
o Media access control (MAC) sublayer:
> Multiple access protocols for channel-access control, for example
CSMA/CD protocols for collision detection and retransmission in
Ethernet bus networks and hub networks, or the CSMA/CA protocol
for collision avoidance in wireless networks.
> Physical addressing (MAC addressing)
2
Link layer tasks in general
o Framing – group bit sequence into packets/frames
> Important: format, size (impact on energy-consumtion and
throughput)
o Error control – make sure that the sent bits arrive and
no other
> Forward and backward error control
o Flow control – ensure that a fast sender does not
overrun its slow(er) receiver
o Link management – discovery and manage links to
neighbors
> Do not use a neighbor at any cost, only if link is good enough
Understand the issues involved in turning the radio
communication between two neighboring nodes into a
somewhat reliable link
3
Error control
o Error control has to ensure that data transport is
>
>
>
>
Error-free – deliver exactly the sent bits/packets
In-sequence – deliver them in the original order
Duplicate-free – and at most once
Loss-free – and at least once
o Error sources: fading, interference, loss of bit
synchronization, …
> Results in bit errors, “bursty” (they tend to occur in clusters
with error-free periods)
> In wireless, sometimes quite high average bit error rates –
10-2 … 10-4 possible!
4
Characteristics of transmission errors
o IEEE 802.11 physical layer frame
>
>
>
MPDU – MAC Protocol Data Unit
Preamble – constant bit pattern for bit and frame synchronization of the receiver
The PHY header describes the lenght and the modulation used in the data part
(MPDU)
o The distortion of waveforms translates into bit errors and packet
losses
o Packet losses (the MAC protocol does not see any data at all)
> The receiver fails to acquire bit/frame synchronization
> The SFD is wrong, the PHY header has an incorrect checksum
o Bit errors
> Synch and PHY header ok, but some bits in MPDU not the same as their
transmitted counterpart
5
Basic approaches to error control
o Backward error control
> Automatic Repeat Request (ARQ)
o Forward error control
> Forward Error Correction
o Different combinations
6
Backward error control – ARQ
o Basic procedure
> Put header information around the
payload
> Compute a checksum and add it to
the packet
• Typically: Cyclic redundancy check
(CRC), quick, low overhead, low
residual error rate
> Provide feedback from receiver to sender
• Send positive (pkt received) or negative acknowledgement (pkt lost)
> Sender uses timer to detect that acknowledgements have not
arrived
• Assumes packet has not arrived
• Optimal timer setting?
> If sender infers that a packet has not been received correctly,
sender can retransmit it
• What is maximum number of retransmission attempts?
7
Standard ARQ protocols
o Alternating bit (or Send-and-Wait)
> Transmitter (XMTR) buffers one packet, sends it and sets a
timer
> RECV receives the packet -> positive ACK -> XMTR frees the
buffer and transmits the next packet
> RECV receives nothing -> negative ACK or being quite -> XMTR
retransmits the packet
> XMTR marks each new packet with sequence numbers
alternating between 0 and 1
> Retransmitted packets are coplies of the original packets
• Have the same sequence number
> The sequence numbers allow the RECV to detect duplicates
• Occurs when positive ACK is lost
8
Standard ARQ protocols
o Goback N
> Send up to N packets, if a packet has not been acknowledged
when timer goes off, retransmit all unacknowledged packets
> Goback N allows the XMTR to have multiple unacknowledged
packets
> XMTR keeps a buffer for up to N packets (a window)
> Each packet has its own timer, started upon the packet’s
transmission
> RECV accepts packets only in sequence and drops packets
received out of sequence (some previous packet had been lost)
• For this RECV needs only buffer space for one packet
> Usually RECV acknowledges the last packet arrieved in sequence
> If at the XMTR the timer for the oldest packet expires before
receiving the corresponding ACK, this packet and all other
packets in the window are retransmitted
9
Standard ARQ protocols
o Selective Repeat
> Similar to Goback N, however here the RECV has N buffers and
uses them to buffer frames arriving out of sequence
> RECV sends either positive or negative ACKs
> XMTR retransmits only those packet for which no ACK has been
received withing the timeout period (more energy-efficient)
o Alternating bit and Selective Repeat retransmit only
erroneous packets (while Goback N potentially
retransmits correctly received packets)
o The number of retransmissions allowed per packet must
be bounded (to avoid spending too much energy)
> A loss-free service therefore cannot be guaranteed
(semireliable protocols)
o Example (Energy consumption of alternating bit)
10
Digital channel model
o How to stochastically capture the behavior of a
wireless channel
o Directly model the resulting bit error behavior
> Each bit is erroneous with constant probability p, independent
of the other bits -> binary symmetric channel (BSC)
11
Example: Energy consumption of alternating bit
12
How to use acknowledgements
o Be careful about ACKs from different layers
> A MAC ACK (e.g., S-MAC) does not necessarily imply buffer
space in the link layer
> On the other hand, having both MAC and link layer ACKs is a
waste
o Do not (necessarily) acknowledge every packet – use
cumulative ACKs
> Tradeoff against buffer space
> Tradeoff against number of negative ACKs to send
13
When to retransmit
o Assuming sender has decided to retransmit a packet –
when to do so?
> In a BSC channel, any time is as good as any
> In fading channels, try to avoid bad channel states – postpone
transmissions
• Retransmissions may be unecessarily lost
> Instead (e.g.): send a packet to another node (if in queue and
if it has better channel) -> exploit multi-user diversity
o How long to wait?
> Example solution: Probing protocol
> Idea: reflect channel state by two protocol modes, “normal”
and “probing”
> When error occurs, go from normal to probing mode
> In probing mode, periodically send short packets
(acknowledged by receiver) – when successful, go to normal
mode
14
Basic approaches to error control
o Backward error control
> Automatic Repeat Request (ARQ)
o Forward error control
> Forward Error Correction
o Different combinations
15
Forward error control
o Idea: Endow symbols in a packet with additional
redundancy to withstand a limited amount of random
permutations
> Additionally: interleaving – change order of symbols to
withstand burst errors (clustered errors)
Source symbols
Information
source
Channel symbols
Channel
encoder
(FEC)
Channel symbols
Interleaver
Digital waveform
Modulator
Tx antenna
Channel
Information
sink
Channel
decoder
Source symbols
Deinterleaver
Channel symbols
Demodulator
Channel symbols
Rx antenna
Digital waveform
16
Forward error control
o Can be used as an open loop technique
> No feedback from the receiver
> Possibly more energy-efficient as no ACK packets are sent
> Moreover, the transmitter does not have to switch from the
transmitting to receiving states in order to be able to receive
ACKs
o In practice FEC is used thogether with acknowlegments
(backward error control) for more robust/reliable
communicaiton
17
Block-coded FEC
o Level of redundancy: blocks of symbols
> Block: k p-ary source symbols (not necessarily just bits)
> Encoded into n q-ary channel symbols
> Mostly, we have p = q = 2, n ≥ k, and the symbols correspond to
bits
o Injective mapping of pk source symbols -> qn channel
symbols
> This mapping is called a code
o Code rate: (k ld p) / (n ld q)
> When p=q=2: k/n is code rate
o For p=q=2: Hamming bound – code can correct up to t
bit errors only if
> Codes for (n,k,t) do not always exist!
18
Hamming distance
o An important metric for a block code
o The Hamming distance of two valid channel words w1 and
w2 is defined as the number of bits in which they differ
> E.g., 1001 and 1011 -> Hamming distance = 1
o The Hamming distance dmin of the whole code is defined
as the minimum Hamming distance of all pairs of valid
channel words
o Any code with a Hamming distance dmin can reliably
detect up to and including dmin − 1 bit errors and can
reliably correct up to and including (dmin−1)/2 bit errors
19
Popular block codes
o Popular examples
> Reed-Solomon codes (RS)
> Bose-Chaudhuri-Hocquenghem codes (BCH)
o Energy consumption
> E.g., BCH encoding: negligible overhead (linear-feedback shift
register)
> BCH decoding: depends on block length (n) and Hamming
distance (that is the number of correctable bits t)
• Eadd and Emult are the energy needed to carry out addition and multiplication
> Similar for RS codes
20
Convolutional codes
o Here also k bits of user data are mapped to n channel symbols; the
coding of two successive k-bit blocks is not independent
o Code rate: ratio of k user bits mapped onto n coded bits
o Constraint length K determines coding gain (by increasing K we can
reduce the transmit power required to achieve fixed BER)
o Energy
> Encoding: cheap
> Decoding: Viterbi algorithm, energy & memory depends exponentially (!)
on constraint length (K)
1
2
k*K
3
Stream of user bits
(k shifted in at once)
……...
Code bits:
+
+
+
Bit 1
Bit 2
Bit 3
……...
+
Bit n
21
Energy consumption of convolutional codes
o Tradeoff between
coding energy and
reduced
transmission power
(coding gain)
o Overall: block
codes tend to be
more energyefficient
22
Interleaving
o Bursy error characteristics affect performances of codes
o Solution: Bit interleaving
>
>
>
An interleaver at the XMTR takes the packet (of multiple blocks) generated by
the FEC encoder
The bits in this packet are permuted before transmitting
The deinterleaver at the RECV inverts the permutation before the packet is
handed to the FES decoder
o The error burst are spread over multiple coding blocks instead of being
concatrated to one or a few blocks
>
Increases the chance that each block can be successfully decoded
Source symbols
Information
source
Channel symbols
Channel
encoder
(FEC)
Channel symbols
Interleaver
Digital waveform
Modulator
Tx antenna
Channel
Information
sink
Channel
decoder
Source symbols
Deinterleaver
Channel symbols
Demodulator
Channel symbols
Rx antenna
Digital waveform
23
Hybrid schemes: FEC and ARQ
o No single fixed error-control strategy will give optimum energy
efficiency at all times
t: error correction capacity
o FEC
o ARQ
> Overhead only when
errors occurred
(expect for ACK,
always needed)
o Both schemes have
their uses ->hybrid
schemes
7
Relative energy consumption
> Constant overhead
for each packet
> Not (easily) possible
to adapt to changing
channel
characteristics
8
6
no FEC
t=2
t=4
t=6
t=8
t=10
5
4
3
2
1
0
1e-07
1e-06
1e-05
0.0001
0.001
0.01
0.1
p
BCH + unlimited number of retransmissions
24
Power control on a link level
o Further controllable parameter: transmission power
> Higher power, lower error rates – less FEC/ARQ necessary
> Lower power, higher error rates – higher FEC necessary
o Tradeoff
o However, by increasing the power of a single XMTR, the
interference on other (neigboring) XMTRs increases ->
i.e., the BER increases as well
> Tradeoffs of this nature are very complex in multihop networks
25
Link layer tasks in general
o Framing – group bit sequence into packets/frames
> Important: format, size (impact on energy-consumtion and
throughput)
o Error control – make sure that the sent bits arrive and
no other
> Forward and backward error control
o Flow control – ensure that a fast sender does not
overrun its slow(er) receiver
o Link management – discovery and manage links to
neighbors
> Do not use a neighbor at any cost, only if link is good enough
26
Framing
o We discuss two aspects of framing
> Choice of packet size
> Intemediate checksum schemes
o The goal is to try to reduce the amount of information
that has to be retransmitted in case of errors while
maintaining minimal overhead
27
Frame, packet size
Energy per useful bit
Energy per useful bit
o Small packets: low
packet error rate, high
packetization overhead
o Large packets: high
packet error rate, low
overhead
o Depends on bit error
rate, energy
consumption per
transmitted bit
o Notation: h(overhead,
payload size, BER)
20
18
16
14
12
10
8
6
4
2
0
1e-05
h(100, 100, p)
h(100, 500, p)
0.0001
0.001
Bit error rate
30
h(100,u,0.001)
25
20
15
10
5
0
0
500 1000 1500 2000 2500 3000
User data size
28
Dynamically adapt frame/packet length
o For known bit error rate (BER), optimal frame length is
easy to determine
o Problem: how to estimate BER?
> Collect channel state information at the receiver (RSSI, FEC
decoder information, …)
> Example: Use number of attempts T required to transmit the
last M packets as an estimator of the packet error rate
(assuming a BSC)
o Second problem: how long are observations valid/how
should they be aged?
> Only recent past is – if anything at all – somewhat credible
29
Intermediate checksum schemes
o With traditional framing the whole packet has to be retransmitted
o In interdediate checksum schemes
>
>
>
The user’s data are partitioned into a number L of chunks, each having c bits to
which a checksum of h bits is appended
The chunks are appended to the original frame header of size m bits
The overall frame has size m+L(c+h) bits
o RECV behaves as follows
>
>
>
>
Error in the header -> discard the whole packet
Check each chunk separatly and buffer the correct ones
If all correct RECV sends final ACK to XMTR
The faulty chunks indicated with and incomplete ACK -> XMTR retransmits only
the faulty packets
o High overhead in low BER scenarios
o Optimal chunk size exists for fixed BER
30
Putting it together: ARQ, FEC, packet length
optimization
o Applying ARQ, FEC (both block and convolutional codes), frame
length optimization to a Rayleigh fading channel
>
Rayleigh fading models assume that the magnitude of a signal that has passed through such a transmission
medium (also called a communications channel) will vary randomly, or fade, according to a Rayleigh distribution
the radial component of the sum of two uncorrelated Gaussian random variables.
31
Link layer tasks in general
o Framing – group bit sequence into packets/frames
> Important: format, size (impact on energy-consumtion and
throughput)
o Error control – make sure that the sent bits arrive and
no other
> Forward and backward error control
o Flow control – ensure that a fast sender does not
overrun its slow(er) receiver
o Link management – discovery and manage links to
neighbors
> Do not use a neighbor at any cost, only if link is good enough
32
Link management
o Goal: decide to which neighbors that are more or less
reachable a link should be established
> Problem: communication quality fluctuates, far away neighbors
can be costly to talk to, error-prone, quality can only be
estimated
o Establish a neighborhood table for each node
> Partially automatically constructed by MAC protocols
33
Link quality characteristics
o Expected: simple, circular
shape of “region of
communication” – not realistic
o Instead:
> Correlation between distance
and loss rate is weak; iso-losslines are not circular but
irregular (see figures)
> Asymmetric links are relatively
frequent (up to 15%)
> Significant short-term Packet
Eerror Rates (PER) variations
even for stationary nodes
Power1
Power2 < Power1
34
Three regions of communication
o Effective region: PER
consistently < 10%
o Transitional region:
anything in between,
with large variation for
nodes at same distance
o Poor region: PER well
beyond 90%
35
Link quality estimation
o How to estimate, on-line, in the field, the actual link quality?
o Requirements
>
>
>
>
Precision – estimator should give the statistically correct result
Agility – estimator should react quickly to changes
Stability – estimator should not be influenced by short aberrations
Efficiency – Active or passive estimator
7
Gap = 2
10
11
Gap = 3
15
Gap = ?
o Example:
EWMA
only estimates
at fixed intervals
EWMA exponentially weighted moving average
36
Conclusion
o Link layer combines traditional mechanisms
> Framing, packet synchronization, flow control
with relatively specific issues
> Careful choice of error control mechanisms – tradeoffs
between FEC & ARQ & transmission power & packet size …
> Link estimation and characterization
37