1

Cross-Layer, Energy-Efficient Design for Supporting Continuous Queries in Wireless Sensor Networks A Quorum-Based Approach

Chia-Hung Tsai, Tsu-Wen Hsu,

Meng-Shiuan Pan, and Yu-Chee Tseng

Springer Netherlands Wireless Personal Communications 2009

2

Outline

Introduction System architecture Quorum layer Query-processing layer Simulation results Conclusions

3

Introduction

Power saving and query processing are two main issues in WSNs

Applying the quorum-based power-saving protocols to the continuous query-processing problem

This paper contributes in proposing a cross-layer approach to integrating the grid-quorum system with continuous queries

4

System architecture

5

Query session information

Each node will maintain a QSI table to keep track of the query paths that currently pass it and the quorums to support these paths

6

Query format

Query is denoted by a 5-tuple (sn, sr, t, p, len) sn is the sink node

sr is the source node t is the lifetime of the query p is the period that sr will generate reports len is the expected packet length per report

7

Quorum layer

Grid quorum system Quorum set for continuous queries

8

Grid quorum system

9

Quorum set for continuous queries The wake-up/sleep schedule of a node will

be determined by one or multiple grid quorums, which we call quorum set

Each grid quorum is denoted by a 4-tupleg = (n1, n2, R, C)

We define the duty cycle of a grid quorum g

10

Example

dty(g) = (4+3-1) / 12 = 1/2

11

Query-processing layer

As more and more continuous queries pass the node, its quorum set will contain more grid quorums

A DSR-like routing protocol will be applied An energy cost function will be defined to

evaluate the quality of a query path

12

Query-processing

Query-Requesting Process Query-Replying Process Query-Removing Process Local Slot Synchronization

13

Query-Requesting Process

Quorum preparing QREQ initiating and processing QREQ rebroadcasting

14

Quorum preparing

When a sink node sn has a query y to a source node sr, it will compute a grid quorum gini to support the query y

y = (sn, sr, t, p, len)

15

Example

y = (sn, sr, t, p, len)

y= (0, 8, 500, 50, 100)

n1=3

n2=3

dty(g) = (3+3-1)/9 = 0.56

len/r * 1/p

100/10 * 1/50 = 0.2

16

QREQ initiating and processing There are two cases involving in producing a

QREQ packet: a node initiates a new query a node receives a QREQ and rebroadcasts it

17

QREQ

We suppose that node xi receives from node xi−1 a QREQ(gini, y, c, PATH) for possibly supporting a query y initiated by node x0

gini is the grid quorum computed by x0

c is the cost calculated by xi−1

PATH is a list of 2-tuples, where each 2-tuple is of the form (node_id, quorum)

18

Duty cycle

xi will find a quorum to serve query y, which we call gser(y)

Given G(xi), we can estimate xi’s duty cycle as follows:

19

Example

1-0.5=0.5

1-0.625=0.375

0.5*0.375=0.1875

Dty=1-0.1875=0.8125

20

Traffic load

From xi’s QSI, we can measure xi’s current traffic load as follows

len(z) is the length of each sensing report p(z) is the period per report for query z

xi’s current traffic load is

21

Capacity(1/2)

xi can measure whether its current quorum set can accommodate y or not by checking LD(xi) + ld(y) ≤ DTY (G(xi))

The capacity of gcan is defined as follows

QS(gcan) means the set of quorum slots of gcan

s-deg(sj) is the share degree of the quorum slot sj in gcan

22

Capacity(2/2)

If there exists one gcan such that

then gcan will be assigned to support y and we will set gser(y) = gcan

23

Costs

The average extra energy cost Cact for xi to remain active per slot

The average extra energy cost Ctx for xi to transmit data for y per slot

24

The average extra energy cost to remain active per slot

Eact is the energy to remain active for one full slot

25

The average extra energy cost to transmit data for y per slot

Etx is the energy to transmit one full slot of data

26

QREQ rebroadcasting

Node xi will also maintain the minimum cost c

min for all paths from x0 to xi that xi has learned so far

27

Query-replaying process

Node xi will collect QREQs for a while and choose the QREQ(gini, y, c, PATH) with the lowest cost c

Then xi will unicast QREP(y, PATH) back to x0

28

Quorum set

If G(xj) = {gdef}, xj will directly set G(xj) = {gser

(y)} Otherwise, xj will set G(xj) = G(xj) {∪ gser(y)} After a node adjusts its quorum set, it can wa

ke up and sleep according to the quorums in its set

29

Query-removing process

Each intermediate node when receiving the QREM(y) will remove the corresponding entry from its QSI table

30

Local slot synchronization

At the clock level, two neighboring nodes will try to synchronize their clocks by aligning their slots

At the quorum level, they will try to synchronize this quorum by aligning the first slot of this quorum at each side

31

Assign priority

Along a query path, a node that is closer to the source node has a higher priority

Between two query paths, the path which was established earlier has a higher priority

32

Simulation results

Set up a 400 × 400 m2 sensing field, on which hundreds of sensor nodes are randomly deployed

Transmission range and carrier sensing range of each sensor node are set to 50 and 100 m

The whole simulation time is 7200 secondsModes transmit receive idle sleep

Energy consumption(mW)

50 45 50 5

33

Quorum setting

The default quorum gdef is set to (40, 40, {1}, {1}) with each quorum slot fixed to 0.1 second

Initially operate under 5% duty cycle and each quorum group is 160 seconds

34

Path sharing

35

Residual energy

36

Impact of our cross-layer design The first one lets each query path adjust its q

uorum on this own [referred to as SP-NC (shortest-path, no-coordination)]

The second one enforces all quorum paths to share the same quorum [referred to SP-GQ (shortest path, global-quorum)]

37

Comparison with the SP-NC Scheme Each query reporting period is set to 60

seconds

38

Comparison with SP-GQ scheme The SP-GQ scheme will pick the quorum with

the lowest duty cycle that can meet all nodes’ requirement as the global quorum

39

Impact of traffic loads

Impact of Transmission Rate Impact of Packet Length Impact of Query Period

40

Impact of Transmission Rate

A smaller transmission rate r will result in slower transmission (and thus a higher traffic load)

Evaluate the energy consumption of our system by varying the transmission rate at 250 kbps, 100 kbps, 50 kbps, and 10kbps

41

42

Impact of Packet Length

Vary the length len per report to evaluate the energy performance

The transmission rate r is fixed to 250 kbps and len varies from 100, 1000, to 5000 bytes

43

44

Impact of Query Period

We set r = 250 kbps and len = 100 bytes and vary the reporting period p from 30 to 70 seconds

A higher reporting period will incur less energy consumption

45

46

Conclusions

Increasing the overlapping of query paths for energy efficiency

We modify the original DSR routing scheme by adding a cost metric to choose quorums along a query path

Simulation results also verify the correctness and performance of the proposed scheme