[PPT] - Outline Wireless Ad Hoc & Sensor Networks Introduction PowerPoint Presentation

SLIDE 1

Wireless Ad Hoc & Sensor Networks

(Wireless Sensor Networks - Part 1)

WS 2010/2011 WS 2010/2011

Prof. Dr. Dieter Hogrefe
Dr. Omar Alfandi

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

2

What are WSNs?

WSNs = Wireless Sensor Networks
Set of individual nodes that are able to interact with the

environment by sensing or controlling physical parameters

Wireless communication enables the cooperation of the

nodes to fulfil bigger tasks that single nodes could not F t Vi i A bi t I t lli

Future Vision: Ambient Intelligence

– Many different devices gather and process information from many different sources to control physical processes and to y p y p interact with human users – WSN is a crucial step towards Ambient Intelligence by providing the “last 100 meters” of pervasive control the last 100 meters of pervasive control

3

Application Examples of WSNs

Disaster relief operations

– Wildfire detection – Sensor nodes with thermometers are dropped from an air plane – Various temperature measurements are Various temperature measurements are collected to produce a temperature map

Biodiversity Mapping

– Gain an understanding about plants and animals

Intelligent Buildings/Bridges

– Measurements about temperature, energy wastage – Monitoring of mechanical stress levels Monitoring of mechanical stress levels

4

SLIDE 2

Application Examples of WSNs

Precision Agriculture

– Precise irrigation and fertilising of fields – Temperature and brightness monitoring

Medicine and health care

P t ti d i t i – Postoperative and intensive care – Long-term surveillance of patients

Logistics
Logistics

– Tracking of parcels during transportation – Inventory tracking in stores or warehouses y g

…and a lot more – almost unlimited possibilities

p

5

Participants in a WSN

sink

Source

– Sensor that senses data in its environment – Can be equipped with different sensors

e.g. temperature, brightness, etc.

– Reports the measurements to the sink Reports the measurements to the sink

Sink (Base Station)

– Interested in receiving data from the other sensor nodes

source

te ested ece g data

t e ot e se so
des

– Can be either part of the WSN or an external device such as a Laptop or PDA I l th i b t ti b t d di th – In general there is one base station, but depending on the application multiple base stations are possible

6

Interaction Patterns between source and sink

Event detection

– If a certain event occurs, the sensor nodes report the t t i t t d i k measurement to interested sinks

Periodic measurement

Periodically reporting of events to interested sinks – Periodically reporting of events to interested sinks

Function approximation and edge detection

– Approximation of a function of space an/or time (e g Approximation of a function of space an/or time (e.g. temperature map) – Edge Detection: Find edges or structures in such a function

Tracking

– Report the position of an observed intruder

7

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks
Types of Source and sinks

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

8

SLIDE 3

Challenges for WSNs – Characteristic requirements

Type of Service

– Not simply moving bits from one place i th t k t th in the network to another – Rather: Provide meaningful information and/or actions about a given task “People want answers, not numbers” (Steven Glaser UC g – Scoping of interactions, e.g.

Geographic regions

Ti I t l

(Steven Glaser, UC Berkeley)

Time Intervals
Quality of Service (QoS)

Traditional QoS metrics do not apply – Traditional QoS metrics do not apply – Adapted quality concepts such as

Reliable detection of events
Approximation quality of a temperature map

9

Challenges for WSNs – Characteristic requirements

Fault tolerance

– Be robust against node failures ( i t f i t f h i l d t ti t ) (running out of energy, interferences, physical destruction etc.)

Lifetime

Normally the replacing of a node energy source is not possible – Normally the replacing of a node energy source is not possible – The WSN should fulfil its task as long as possible  energy-efficient operation – Trade-off: Lifetime vs. QoS – BUT: What is the precise definition of Lifetime?

Ti til fi t d f il

Time until first node fails
50% of nodes failed
Certain geographical area is not covered anymore

 Not uniquely defined!

10

Challenges for WSNs – Characteristic requirements

Scalability

– Architecture and protocols need to support a large amount of sensors

Wide range of densities

Number of nodes per unit area differs  application dependent – Number of nodes per unit area differs  application dependent – Change over time due to node movement or node failures

Programmability

Programmability

– Increase the flexibility by enabling the re-programming of nodes in the field to react to new situations

Maintainability

– Environment and WSN itself are changing  self monitoring and adaptation of the system  self-monitoring and adaptation of the system

11

Challenges for WSNs – Required mechanisms

Multihop wireless communication

– To save energy limit radio range – Use intermediate nodes as relays

Energy-efficient operation

S i t ti d i ti – Sensing, computation and communication

Auto-configuration

For a huge amount of sensors manual configuration is no option – For a huge amount of sensors manual configuration is no option

Collaboration and in-network processing

– Node collaborate to achieve a common goal Node collaborate to achieve a common goal – To improve efficiency the sensed data can be aggregated  e.g. calculation of the average temperature

12

SLIDE 4

Challenges for WSNs – Required mechanisms

Data-centric networking

– Focus on relevant data, not on the node which is providing it  “R i l if t t d 30°C”  e.g. “Raise an alarm if temperature exceeds 30°C” – Nodes are characterised by the provided data (data-centric), not by the network address (address-centric) y ( )

Locality

– Do thing locally as far as possible, i.e. on the node itself or in collaboration with its neighbours

Exploit trade-offs

M t ll t di t l – Mutually contradictory goals – e.g. Energy vs. accuracy

13

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

14

Why are WSNs different? – WSNs vs. MANETs

WSN MANET Applications and equipment

Small sensor nodes with

constrained hardware and

Powerful nodes (laptop, PDA)

with large batteries q p energy supply

In general, unattended
peration

g

In general, more elaborate

applications, e.g. VoIP, with human interaction Application specific

Infinite number of

applications in terms of

Although, a few scenarios not

as many as in WSNs devices, protocols, density etc. Environment

Lot of environmental
More conventional human-

Environment interaction

Lot of environmental

interactions

low data rates, but also data

bursts  new traffic patterns

More conventional human-

driven applications with well- understood traffic characteristics bursts  new traffic patterns characteristics

15

Why are WSNs different? – WSNs vs. MANETs

WSN MANET Scale

Huge amount of sensor nodes

 more scalable solutions

Significantly less nodes than

in WSNs required (e.g. Protocols without node identifiers) Energy

Tighter requirements mostly
Energy constrained but

Energy Tighter requirements, mostly no recharge or replacement of batteries possible Energy constrained, but

ften energy can be

recharged Self

Almost equal to MANETs but
One of the main features in

Self configurability

Almost equal to MANETs, but

different data traffic and energy trade-offs

One of the main features in

MANETs D d bili I di id l d i i l E h d h ld b li bl Dependability and QoS

Individual node is irrelevant as

long as network is working

New QoS concepts necessary
Each node should be reliable
QoS determined by

applications such as VoIP jitt jitter

16

SLIDE 5

Why are WSNs different? – WSNs vs. MANETs

WSN MANET Data centric

Redundant deployment

makes data centric protocols

Data-centric protocols are

more or less irrelevant for p attractive MANETs Simplicity and resource

OS and software must be

simpler than on ‘normal’ PCs

Slightly limited resources, but

in general normal OS and resource scarceness simpler than on normal PCs

Breaking of strict network

layers isolation to achieve simplicity in general normal OS and applications can run on the nodes simplicity Mobility

Mostly stationary use, but

movement for certain applications possible e g

One of the main features of

MANETs  caused by moving nodes applications possible e.g. tracking applications

Movement can be correlated

e g sensors carried by a river nodes

Movement can be correlated

by moving groups e.g. sensors carried by a river

17

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

18

Basic Node Architecture

Controller

– processes all l t d t

Controller Memory Sensor(s) Communication

relevant data – capable of executing arbitrary code

Controller Sensor(s) Power Supply Device

y

Memory

– stores data and programs often different types are used

Supply

Communication

– Device for sending and receiving data over a wireless channel

Power Supply

– Some form of batteries to provide energy; sometime recharging by obtaining energy from the environment e g solar cells by obtaining energy from the environment, e.g. solar cells

19

Basic Node Architecture – Controller

For the sensor nodes mostly microcontroller are used

– general purpose processor, optimised for embedded li ti l ti applications, low power consumption

Examples

Intel StrongARM – Intel StrongARM

High-end processor often used in PDAs

– SA-1100 model: 32-bit RISC core, running @206MHz

– Texas Instruments MSP 430

Intended for usage in embedded applications

– 16-bit RISC core, up to 4 MHz, 2-10 kB RAM, several DACs, RT clock 16 bit RISC core, up to 4 MHz, 2 10 kB RAM, several DACs, RT clock

– Atmel ATMega

ATMega 128L: Intended for usage in embedded applications

8 bit t ll l th MSP430 b t l – 8-bit controller, larger memory than MSP430, but slower

20

SLIDE 6

Basic Node Architecture – Memory

RAM (Random Access Memory)

– To store intermediate sensor readings, k t f f di t packets for forwarding etc. – Is fast, but looses content if power supply is interrupted p

ROM (Read-Only Memory)

– To store fix programs; not writeable

EEPROM or Flash Memory

– Enables overwriting of data – Can be used as intermediate storage if RAM is insufficient or if the RAM’s power supply should be turned-off – BUT: long read/write access delays high energy requirement BUT: long read/write access delays, high energy requirement

21

Basic Node Architecture – Communication

Transceivers

– Both a transmitter and a receiver are required in a sensor node – For practical purposes these two tasks are often combined in one entity, the so called transceiver – Usually half-duplex operation is used because transmitting and Usually half duplex operation is used, because transmitting and receiving at the same time is impractical in the wireless medium

Transceiver States

– Transmit – Receive Idl (R d t i b t tl i i ) – Idle (Ready to receive, but currently no receiving)

Some functions in the hardware can be turned-off to save energy

– Sleep (Significant parts of the receiver are switched off) p ( g p )

No receiving; Recovery time and start-up energy must be considered

22

Basic Node Architecture – Communication

Transceiver characteristics

– Capabilities

Interface: bit, byte or packet level?
Frequency range? (typically: 2.4 GHz, ISM band)
Multiple Channels? Data Rate? Range?

p g

– Energy Characteristics

Power consumption for sending and receiving data?

Ti d ti t h b t t t ?

Time and energy consumption to change between states?
Transmission power control? Power Efficiency?

– Radio Performance

Modulation (ASK, PSK, …)? Noise figure (NF=SNRI/SNRO)?
Gain (signal amplification)? Receiver Sensitivity? Blocking

Perfomance? Out of band emissions? Carrier sense and RSSI? Perfomance? Out of band emissions? Carrier sense and RSSI? Frequency stability? Voltage range?

23

Basic Node Architecture – Power Supply

Goal

– Provide as much energy as possible at smallest cost/ volume/ weight/ recharge time cost/ volume/ weight/ recharge time

Options

– Primary Batteries (not rechargeable) – Secondary Batteries (rechargeable)

Requirements

C it (hi h t ll i ht ll l l i ) – Capacity (high at small weight, small volume, low price) – Capacity under load (withstand various usage patterns) – Self-discharge (if low  long lifetime) g ( g ) – Efficient recharging (at low current; no ‘memory effect’) – Relaxation (exploit the ‘self-recharging effect’) Voltage stability (DC DC conversion) – Voltage stability (DC-DC conversion)

24

SLIDE 7

Basic Node Architecture – Power Supply

Primary Batteries Chemistry Zinc-Air Lithium Alkaline E (J/

3)

3780 2880 1200 Energy (J/cm3) 3780 2880 1200 Secondary Batteries Secondary Batteries Chemistry Lithium NiMHd NiCd Energy (J/cm3) 1080 860 650

25

Basic Node Architecture – Energy Scavenging

Batteries can often not be recharged on the traditional way

 Therefore: Scavenging of energy from the environment

Energy scavenging approaches

– Photovoltaic (solar cells)

b t 10 W/

2

d 15 W/

2

between 10 µW/cm2 and 15 mW/cm2

– Temperature gradient

~ 80 µW/cm2 @ 1V from 5K difference

µ @

– Vibration

between 0.1 and 10.000 µW/cm3

P i ti ( i l t i ) – Pressure variation (piezo-electric)

~ 330 µW/cm2 from the heel of a shoe

– Flow of air/liquid Flow of air/liquid

MEMS gas turbines  no real results yet

26

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

27

Operating Systems for WSNs

Usual Tasks of an Operating System (OS)

– Controlling and protecting the access to resources and managing their allocation to different users their allocation to different users – Support for concurrent execution of several processes and communication between these processed

Disadvantage of running a traditional OS on sensor nodes

– Tasks are only partially required in embedded systems  code is restricted  code is restricted – Microcontrollers do not have the required resources to run a full blown OS – Special requirements such as energy-efficient execution, energy management etc. are not supported – Efficient handling of multiple (asynchronous) external Sources is g p ( y ) mostly not supported

28

SLIDE 8

Operating Systems for WSNs – Concurrency

Traditional concurrency

approach: Processes/ Threads

Handle sensor process Handle packet process

– Based on interrupts, context switching

BUT: Process/ Thread approach
BUT: Process/ Thread approach

is not suitable for WSNs

– One process per protocol leads to p p p too many context switches – Memory and execution overhead is too big for available memory too big for available memory – Mostly only ‘little’ tasks so that expensive context switching is not

OS-mediated

justifiable

OS mediated process switching 29

Operating Systems for WSNs – Concurrency

Better approach for WSNs:

event-based programming

Sensor event Radio event

– Perform regular processing

r be idle

– React to emerging events

Radio event handler Sensor event handler Idle/ Regular processing

React to emerging events immediately when they happen – Basically: Interrupt handler – Normally two contexts

1. First context for time critical event-handlers

 event-handlers are simple and short p  the execution of an event-handler cannot be interrupted  event-handlers are required to run to completion

2. Second context for processing normal code

p g

 performance improvement, memory and energy reduction

30

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

31

Famous sensor nodes

Mica Mote

– Family of sensor nodes, started in th l t 1990 the late 1990s – University of Berkeley, Cooperation with Intel

EYES nodes
EYES nodes

– Developed by Infineon – Sponsored by the EU-project “EYES” p y p j

BTnodes

– Developed at the ETH Zürich

Scatterweb

– Developed at the Computer Systems & T l ti t th FU B li Telematics group at the FU Berlin

32

SLIDE 9

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

33

Types of source and sink

Sources

– any entity that provides data/ measurements – typically a sensor node or an actuator node given feedback

Sinks

d h i f ti i i d – nodes where information is required – Three sorts of sinks

sink is a sensor/ actuator

sink is a sensor/ actuator

sink is a device such as a PDA
sink is a gateway to another network such as the Internet
Multiple Sources and/ or sinks are also possible

– here: mostly one sink and multiple sources are considered

34

Single-hop vs. multihop networks

Common problem: Limited range of wireless communication

– Due to limited transmission power, path loss, obstacles etc.

Option: Multihop Networks

– Send packets to intermediate node I t di t d f d k t – Intermediate node forwards packets in the direction of the destination – Store-and-Forward multihop network p – Basic technique that is used in MANETs and WSNs

N t

bstacle
Note:

– There are other techniques such as collaborative networking and network coding which are not considered here network coding which are not considered here

35

Energy efficiency of multihop routing

Intuitive approach

– Attenuation of radio signals is at least quadratic in most i t environments – It consumes less energy to use relays instead of direct communication – Radiated energy required for distance d is reduced from cdα to 2c(d/2)α (c some constant, path loss α > 2)

H thi h i i lifi d

However, this approach is over-simplified

– Only radiated energy is considered, but the actual consumed energy

Number one myth of multi-hopping: it saves energy

but the actual consumed energy (particularly of the relay nodes) is omitted – Great care should be taken when applying lti h i ith th l f i i ffi i

(Min and Chandrakasan)

multi-hopping with the goal of improving energy efficiency

36

SLIDE 10

Node Mobility

Node mobility

– Sensor nodes itself are mobile Application dependent e g environmental – Application dependent, e.g. environmental control (static) vs. livestock surveillance (mobile) – Deliberately, self-driven vs. driven by external force – Targeted vs. random movement

Sink Mobility

Information sink that is not part of the WSN e g user with PDA – Information sink that is not part of the WSN, e.g. user with PDA – Mobile requester

Event Mobility

– Cause of events or the object that should be tracked moves – Different sensors, which are in the range of the object/event, are responsible for the surveillance responsible for the surveillance

37

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

38

WSN Optimisation Goals

Different scenarios, application types and network

solutions  Ch ll i ti  Challenging questions:

– How to optimise a network? How to compare these solutions? – How to compare these solutions? – Which approach gives better support for my application?

General answers appear impossible, but a few aspects

General answers appear impossible, but a few aspects should be considered

– Quality of Service – Energy Efficiency – Scalability R b t – Robustness

39

WSN Optimisation Goal – Quality of Service

QoS in MANET

– Low-level QoS:

Throughput, delay, jitter, packet loss

– High-level QoS

Perceived QoS e.g. for multimedia applications

Perceived QoS e.g. for multimedia applications

QoS in WSNs is more complicated

– Event detection/ reporting probability p g p y – Event classification error – Detection delay – Probability of missing a periodic report – Approximation accuracy Tracking accuracy – Tracking accuracy

40

SLIDE 11

WSN Optimisation Goal – Energy efficiency

“Energy efficiency” covers several aspects

– Energy per correctly received bit – Energy per reported (unique) event – Delay/energy trade-offs Network lifetime – Network lifetime

Time to first node death
Network half-life
Time to partition
Time to loss of coverage
Time to failure of first event notification

Time to failure of first event notification

The evaluation of these metrics needs clear assumptions

– Node’s energy consumption, network load, radio channel behaviour gy p , ,

41

WSN Optimisation Goal – Scalability

Maintain performance characteristics

regardless of the number of nodes

Typical node numbers

– MANETs: up to hundred nodes WSN th d f d – WSNs: thousands of sensor nodes

Information that have to be globally consistent should be

minimised due to resource limitations e g limited memory minimised due to resource limitations e.g. limited memory

Scalability has direct consequences for the protocol

design design

– Often penalty in performance and/or complexity for small networks  implement appropriate scalability support for your specific application rather than trying to be as scalable as possible

42

WSN Optimisation Goal – Robustness

Robustness is directly related to QoS and scalability

requirements

WSNS should not fail just because

– a limited number of nodes run out of energy h i th i t – changes in the environment – interruption of radio links etc.

Failures have to be compensated
Failures have to be compensated,

e.g. by finding an alternative route, if a link breaks down

Precise evaluation of robustness is difficult

Precise evaluation of robustness is difficult

– depends mostly on failure models for nodes and communication links in practise

43

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

44

SLIDE 12

Design principles for WSNs

The discussed optimisation goals does not provide any

hints how they can be achieved by the network structure

Some basic principles for designing network protocols

– Distributed organisation I t k i – In-network processing – Adaptive fidelity and accuracy – Data centricity Data centricity – Exploitation of

Location information
Activity patterns
Heterogeneity

– Component-based protocol stacks and cross-layer optimisation Component based protocol stacks and cross layer optimisation

45

Design principle – Distributed organisation

WSN nodes should cooperatively organise the network

using distributed algorithms and protocols  lf i ti  self-organisation

Potential shortcomings

Oft t li d h d l ti th t f – Often a centralised approach can produce solutions that perform better and use less energy, when taking all overheads into account

Option: “limited centralised” solution

– Elect nodes for local coordination/ control  ti hi h  creating a hierarchy – Perhaps rotate this function over time

46

Design principle – In-network processing

Nodes in the network are actively involved in taking

decisions about how the network should operate  BUT li it d t i f ti b t th t k it lf  BUT: limited to information about the network itself

This approach can be extended to information

processing processing

– arbitrary extensions providing any form of data processing to improve the application p g p pp

In-network processing techniques

– Aggregation – Distributed source coding and distributed compression – Distributed and collaborative signal processing M bil d / t b d t ki – Mobile code/ agent-based networking

47

Design principle – In-network processing II

Example: Aggregation

– Application of composable aggregation f ti t t t functions to a converge cast tree – Typical functions:

Minimum, maximum, average, …

1.42 , , g ,

– Information is aggregated into a condensed form and then transmitted  t ffi i

1.5 2 5 1

 greater energy-efficiency

1.5 2 2.5 Example: Average Example: Average

48

SLIDE 13

Design principle – Adaptive fidelity

Adapt the effort for exchanging data to the currently

required accuracy/ fidelity  save energy

Example: event detection

– When there is no event, only send rarely an “I am alive”-message If t i th t f – If event occurs, increase the rate of messages

Example: temperature monitoring

When temperature is in an acceptable range only send – When temperature is in an acceptable range, only send temperature values at low resolution – If temperature exceeds a certain threshold, increase the resolution

49

Design principle – Data centric networking

In typical networks, including MANETs, transactions are

addressed to the identity of certain nodes

– “node-centric” or “address-centric” networking paradigm

However, in WSNs sensors are redundantly deployed,

the specific source of an event is not important

nly the

the specific source of an event is not important – only the data itself is relevant

Thus focus on the data of networking transactions
Thus, focus on the data of networking transactions,

instead of the senders and receivers

– “data-centric” networking paradigm g p g – Decoupling of identities and decoupling of time

50

Design principle – Data centric networking II

Implementation options for data-centric networking

– Overlay networks and distributed hash tables (DHT)

Data is stored in a table and is identified via a given key (hash)
DHT will provide one or multiple sources for the data associated

with the key  efficient data source lookup

WSN challenges

– Static keys in DHT vs. dynamic requests in WSNs – DHT ignores distance/ hop count between nodes, i.e. adjacent g p , j neighbours are defined on the basis of semantic information

– Publish/ Subscribe

Nodes can publish data; Nodes can subscribe for any kind of data
Nodes can publish data; Nodes can subscribe for any kind of data
If data is published, the data will be delivered to all subscribers

– Database

Querying the database for certain aspects, e.g. in SQL

51

Design principle – Exploitation

Exploitation of location information

– Location of an event is a crucial information for many applications d it th f l d il bl and it therefore, already available – Exploit this information to simplify the design and operation of communication protocols; improve energy efficiency p ; p gy y

Exploitation of activity patterns

– Protocol Design should consider special activity patterns in WSNs such as low activity and when an event occurs high activity, i.e. bursts of traffic

Exploitation of heterogeneity
Exploitation of heterogeneity

– By construction: different types of nodes in the network – By evolution: some nodes have higher workload, thus less energy y g , gy

52

SLIDE 14

Design principle – Cross-layer optimisation

Normally the network layers are strictly separated

– The functionality of each layer is implemented in a component – Certain interfaces are provided to enable communication between the layers

In WSNs sometimes these strict lines are passed
In WSNs sometimes these strict lines are passed

(Contrary to the rules of standard networking!)

– Big Performance gain g g – However, new problems occur

Feedback loops

E d i f ti lit d f f th ti t

Endangering functionality and performance of the entire system

53

Outline

Introduction
Challenges in WSNs
Differences MANET vs. WSN
Basic Node Architecture

O ti S t f WSN

Operating Systems for WSNs
Famous Sensor nodes
Types of Source and sinks/
Types of Source and sinks/

Single-hop vs. Multi-hop/ Node mobility

WSN optimisation goals

WSN optimisation goals

WSN design principles
Summary

54

Summary

WSN have (almost) unlimited application space
There are a lot of special characteristics that have to be

considered for WSNs, which differ from common networks

Although MANET and WSNs have a similar background

there are several differences Th d hit t ll th ti t

The node architecture as well as the operating system

have to be specially adapted to the application area of the WSN the WSN

There are several optimisation goals and design

principles that should be taken into account for p p developing and implementing a WSN

55