Wireless Networks of Device (WIND) Hari Balakrishnan and John - - PowerPoint PPT Presentation

wireless networks of device wind
SMART_READER_LITE
LIVE PREVIEW

Wireless Networks of Device (WIND) Hari Balakrishnan and John - - PowerPoint PPT Presentation

Wireless Networks of Device (WIND) Hari Balakrishnan and John Guttag MIT Lab for Computer Scienc e http://wind. lcs . mit . edu/ NTT-MIT Meeting, January 2000 Motivation Networks today Future networks Clients Audio Video Sensors


slide-1
SLIDE 1

Wireless Networks of Device (WIND)

Hari Balakrishnan and John Guttag MIT Lab for Computer Scienc e http://wind. lcs . mit . edu/

NTT-MIT Meeting, January 2000

slide-2
SLIDE 2

Motivation

Routers Web servers Clients

Networks today Future networks

Sensors Mobile users Audio Video Devices Wireless

slide-3
SLIDE 3

Location-based wireless servi

  • Communication & control
  • Mobility & group

communication

  • Spontaneous networking
  • Automatically obtain act

map of region

slide-4
SLIDE 4

WIND project

  • Configuration

– Topology formation in mobile networks

  • Routing

– Protocols for mobile device netw

  • rks
  • Resource discovery

– New naming system for dev ice networks

  • Adaptation

– Congestion Manager for multimedia

  • Security & privacy

– Private location support system

slide-5
SLIDE 5

Resource discovery

  • Problem: advertise and locate netwo

rked services

  • Heterogeneous environment

– Wide variety of devices, services, and information

  • Dynamic environment

– Mobile nodes and services – Variable, unpredictable performance – Device failures – Services composed of gro ups of nodes

slide-6
SLIDE 6

Responsiveness Integrate name resoluti and message routing to handle mobility Robustness Easy configuration Name resolvers self- configure into overlay network Expressiveness Decentralized, cooperat resolvers

Solution: Intentional Namin System (INS)

Names are intentional; know what , not where

slide-7
SLIDE 7

Naming and service discover

  • Wide-area naming (DNS)
  • Attribute-based systems

(X.500)

  • Brokers (CORBA

, ANSA)

  • Service location protocol (IETF)
  • Device discovery

– Jini , Universal plug-and-play

  • Intentional Naming System (INS)

– Mobility & dynamism via late binding – Decentralized, serverless

  • peration

– Easy configuration

slide-8
SLIDE 8

Intentional names

[vspace = mit . edu/thermomet er] [building = ne43 [floor = 5 [room = *]] [temperature < 60

0F]

data [vspace = lcs .mit. edu/camera] [building = ne43 [room = 510]] [resolution=800x600] [access = public] [status = ready]

  • Expressive name language (like XML)
  • Providers announce descriptive names
  • Clients make queries

– Attribute-value matches – Wildcard matches – Ranges

slide-9
SLIDE 9

INS architecture

Name Name resolver resolver

Overlay netw

  • rk of

resolvers

Client Client Name Service Service

[building = ne-43 [room = 510]] [entity = camer a]

Query Periodic advertisement

slide-10
SLIDE 10

Name resolver functions

  • Name lookup
  • Routing protocol
  • Handling service mobility
  • Message forwarding
  • Self-configuration algorithm
slide-11
SLIDE 11

Routing protocol tracks chan

Name Name resolver resolver

Overlay netw

  • rk of

resolvers

Client Client Name Service Service

[building = ne-43 [room = 510]] [entity = camer a]

Triggered update Service mobility

[building = ne-43 [room = 510]] [entity = camer a]

slide-12
SLIDE 12

Late binding handles mobili

Name Name resolver resolver

[building = ne-43 [room = *]] [entity = camer a] flag = ANY

data

[building = ne-43 [room = 510]] [entity = camer a] [building = ne-43 [room = 504]] [entity = camer a]

Forward to best location Intentional anycast

slide-13
SLIDE 13

Intentional multicast for gr communication

Name Name resolver resolver

[building = ne-43 [room = *]] [entity = camer a] flag = ALL

data

[building = ne-43 [room = 510]] [entity = camer a] [building = ne-43 [room = 504]] [entity = camer a]

Forward along spanning tree

slide-14
SLIDE 14

Two notable properties

  • Robustness

– Distributed cooperation without centralize servers – Routing information about names – Multicast forwarding

  • Self-configuration

– Resolver network topology based on networ k performance – Decentralized spanning tree algorithm

slide-15
SLIDE 15

Status

  • Java implementation of INS & applications

– Active map, video delivery, audio jukebox, camera network (de mo) – Alpha code release for NTT

  • Scalability & deployment

– Wide-area architecture being des igned – Standardization: MIME for devices/serv ices (XML-based)

slide-16
SLIDE 16

Summary

  • Future networks

will include heterogeneous devi & services – Configuration, routing, discovery, adaptat security

  • INS handles resource discove

ry in dynamic, mobile networks – Expressiveness: intentional names – Responsiveness: late binding – Robustness: decentralized resolvers – Configuration: self-configuration protocol

http://wind. lcs . mit . edu/