• Aucun résultat trouvé

Middleware to Support Sensor Network

N/A
N/A
Protected

Academic year: 2022

Partager "Middleware to Support Sensor Network "

Copied!
13
0
0

Texte intégral

(1)

Middleware to Support Sensor Network

Applications

Windi B.Heinzelman,Amy L.Murphy,hervaldo S.Carvalho,Mark A.Perillo

University of Rochester

IEEE Network. January/February2004 IEEE Network. January/February2004

(2)

Plan

Introduction

Sensor Network Management

Problems

Existing solution

Relation among different middleware

MiLAN

How MiLAN works?

conclusion

(3)

Introduction

What is Sensor Network?

App 1

App n Network

Network M

M

M M M

System QoS QoS

QoS Data

Data

Middleware (MiLAN)

(4)

Sensor Network Management

Applications: they need a specific QoS

Sensors:

Distributed

Energy limited

Channel Bandwidth

Problems

(5)

Sensor Network Management

Existing solution: protocols to extend network lifetime

Low-level node collaboration (LEACH, Directed Diffusion)

Turn nodes off whenever possible (PAMAS, S-MAC, …)

Tailoring the routing protocols to the characteristics of sensor networks (Rumor Routing,SPIN)

(6)

Relation among different middleware

MiLAN

(7)

MILAN

(Middleware Linking Application & Networks)

Application/captures

MiLAN

Plugin Abstraction Milan commands

Convert

What makes it different?

(8)

How MiLAN works?

fi fA

fN

Sensor QoS

variable State Requiremnts

Sensor(s) MiLAN Network Plug-in

MiLAN Application(s)

C B

A

fN

state

Data Node discovery

:Application feasible set :Network feasible set :Best performance/cost trade-off

fN

Configure

fi fi

Determine Network type

Calculate fA

Calculate

fN

Tradeoff Among

fN

fA I

(9)

Scope of MiLAN applications

Environmental surveillance

Home/Office security

Medical Monitoring

(10)

Example about the utilization of MiLAN

Sensor-based personal health monitor

None Low High

User stress state Med System state

Resp. rate Blood O2

.3 .8 .3 variables

Norm High/low Respiratory rate

Norm High/

Low

High/

[+ECG]

Heart rate Norm Low

Blood pressure Var.

state

.7

Blood flow

Blood press ECG

1.0 .8 .7

sensors Pulseoxy

1.0

Virtual sensor

(11)

Conclusion

Middleware eases the application development task in complex environment.

The needs of the application should be

integrated with the management of the network in a single middleware system

Trade the application performance for network cost

(12)

Personal critics…

The article doesn’t discuss the pervasive environment…

Lack of a case study about MiLAN’s

application

(13)

For more informations

MiLAN Project:

http://www.futurehealth.rochester.edu/milan/

Références

Documents relatifs

Density of congestion window increase (α) parameters chosen by intra-layer cognitive engine. Results confirm that the proposed cognitive adaptation leads to significant

It makes the Border Gateway Protocol (BGP) hijacking attack [2] possible in the Bitcoin network, where an attacker can delay the information propagation and partition the

In this paper, we realize energy efficient and QoS based network layer which uses RRR routing protocol for security management of chemical products warehouse, and we evaluate his

The model of capture is called binary due to the fact that a frame is captured or not according to the received powers at the receiver from concurrent frames and depends to the

In the order to demonstrate the performance of our new algorithm, we will compare it to three methods, namely the EP method (Equal Power) in which we attribute the power to each

Perspectives of the mobile robot middleware and application development in the context of wireless sensor networks are mainly focused on issues of deploy- ment algorithm

Fig. 8: Effect of total dose on the defect cluster size distribution in self-ion irradiated tungsten, at a dose rate of 0.0125 dpa/s and a maximum mobile SIA cluster size of N i max

From the total water content evolution in time, the dynamic and static water contributions were dissociated using the values of water trapped in the solid waste bed after each