Context Aggregation and Dissemination in Ubiquitous Computing Systems

1
Context Aggregation and Dissemination in
Ubiquitous Computing Systems

• 2003. 9. 9
• Kiwon Hong
• (kiwon77.chonnam.ac.kr)

2
contents

• Introduction
• Motivation
• The abstraction
• Summary

3
Introduction

• Ubiquitous computing environment
• user may interact with hundreds of devices
• user attention becomes a scarce resource
• application(context-aware)
• need a constant flow of information
• respond to context changes by adapting to the new
  context
• must be aware of the context in which they run
• can reduce user distraction by dynamically
  adjusting their behaviors to the current context
• current state of the user, computational
  environment, physical environment

4
Introduction

• Context information
• be derived from an array of diverse information
  sources
• location sensors, weather or traffic sensors, and
  the status of computational or human services

5
Motivation
operator
Application-specific processing
6
motivation

• the burden of converting source data into context
  information
• is on servers in the network, not on application
  platforms
• the decomposed graph structure
• improve flexibility, scalability
• avoiding the transmission of unnecessary data to
  application

7
the abstraction

• Propose an abstraction for context collection,
  aggregation, dissemination based on a directed
  acyclic graph(DAG)
• event
• context information
• information source
• sensor
• event stream
• The sequence of event
• publisher
• produces an event stream
• subscriber
• consumes an event stream
• operator
• an object that subscribes to and processes on or
  more input event stream and publishes an other
  event stream

8
the abstraction

• operator graph
• An event-flow graph
• composition sources, operators, and applications
• sources
• no subscriptions
• wrappers for context sensors
• operator
• deterministic functions of their input events
• application
• sinks of the graph
• react to incoming events
• may not be a tree -gt multiple input , output
  event
• publishers source, operator
• subscribers operator, application

9
the abstraction

• categories of operators
• filter
• outputs a subset of its input events
• transformer
• only changes some attribute values
• merger
• outputs every event it receives
• aids re-use of event stream
• aggregator
• outputs an arbitrary type event stream based on
• the event in one or more input stream event
  stream

10
the abstraction
11
summary

• Directed Acyclic Graph (DAG)
• the contextual information sources as event
  publishers
• applications do not have unique needs
• to share some of the processing between
  applications and users -gt DAG
• offers flexibility and scalability
• flexibility
• composition of generic and custom operators
• scalability
• re-use of event stream