ContextAwareness and Coordination in Mobile Environments

1
Context-Awareness and Coordination in Mobile
Environments

• Gruia-Catalin Roman
• 1 April 2003
• Mobile Computing Laboratory
• Department of Computer Science and Engineering

2
Project Team

• Faculty
• Gruia-Catalin Roman
• External Collaborators
• Doctoral Students
• Chien-Liang Fok
• Radu Handorean
• Qingfeng Huang
• Christine Julien
• Jamie Payton
• Rohan Sen

• Masters Students
• Yigit Cakar
• Randy Pitz
• Undergraduate Students
• Greg Hackmann
• Shannon Lieberg

3
Project Profile

• Research goals
• Rapid development of dependable mobile
  applications over ad hoc networks
• Ad hoc network challenge
• No fixed network infrastructure
• Frequent disconnections
• Transient interactions
• Decoupled computing
• Limited guarantees
• Vulnerability

4
Project Profile

• Computational model
• Physically mobile hosts
• containers equipped with wireless connectivity
• Logically mobile agents
• units of mobility, modularity, and execution
• Agent data
• tuple spaces and more
• Solution strategy
• Focus on adaptation mechanisms
• Reliance on context-awareness
• Convergence around coordination models
• Formal specification and analysis

5
Project Profile

• Technology transfer
• Models made manifest as middleware
• Industrial collaborations (Ford)
• Evaluation scope
• Wide range of application types

6
Range of Research Activities

• Formal models of mobility
• Mobile UNITY
• Coordination models for mobility
• Global Virtual Data Structures
• Middleware for mobility
• LIME Limone
• Algorithms and protocols for mobile computing
• Mobile applications

7
Current Activities

• Spatiotemporal services and analysis methods in
  mobile and sensor networks
• Lightweight coordination among resource
  constrained devices
• Secure service provision in ad hoc networks
• Wide-ranging exploration of context-awareness
• Asymmetric models
• Extended scope
• Declarative specification
• Transparent maintenance

8
Spatiotemporal Services
Send data to delivery zone Z(t)
Z(1)
Z(2)
Z(3)
Mobicastsource
9
Lightweight Coordination
Host 1
Reaction Fired
Migration
Host 3
Host 2
Operation Request
10
Secure Service Provision
application
application
service provision
service provision
secure tuples secure tuple spaces
secure tuples secure tuple spaces
security table
security table
L I M E
L I M E
remote interactions
interceptor
interceptor
11
Network Abstractions for Context-Aware Mobile
Computing

• Christine Julien
• 1 April 2003
• Mobile Computing Laboratory
• Department of Computer Science and Engineering

12
Motivating Application
13
Context-Awareness for Adaptability in Ad Hoc
Networks

• Continuous and rapid reaction to changes
• Application specific context definition
• Explicit controls over the scope size and cost
• Multitude of contexts evolving over time
• Scope that extends beyond the local host
• Generalized interaction with context types

14
Simplifying Application Development

• Provide generalized view of context extended to
  include information about distant hosts
• Provide more flexibility to applications by
  allowing application specific context
  specifications
• Provide transparent maintenance of these contexts
  through automatic context sensing
• Help programmer cope with problems inherent in ad
  hoc networks

15
Solution Strategy

• Allow multiple contexts that change over time
• Expand contexts to encompass a surrounding
  neighborhood
• Provide declarative specifications of these
  contexts
• Encompass the specific needs of an application
• Control the scope and size of the view
• Provide varying perceptions of context
• Provide a protocol for calculating contexts
• Middleware for rapid mobile application
  development

16
Network Abstraction

• Acquaintance List
• Set of nodes that contribute information used to
  build a context for a reference node
• Network Abstraction
• Allows specification of the list
• Extends availability of context information
• Limits scope of operation over network

17
Solution Sketch

• Represent the ad hoc network as a graph
• Abstract properties of nodes and links to weights
  on edges in this graph
• Calculate the cost of paths from the reference
  node based on a general cost function
• Determine the shortest possible path to each
  node, and build a tree of these paths
• Limit this tree using a bound D

18
Weight Assignment

• Individualized for a particular application
• Host properties combine to form ?i
• ?i battery power, CPU power, physical location,
  etc.
• Link properties combine to form ?ij
• ?ij physical distance, bandwidth, throughput,
  etc.
• Each links logical weight combines link
  properties and node properties
• mij ?(?i, ?j, ?ij)

19
Computing Path Cost

• Assuming each link has a single weight, an
  application can define a cost function
• v0(Pk) Cost(v0(Pk-1), mk-1,k)
• v0(?v0?) 0
• Cost function must be strictly increasing along a
  path
• Sample metrics
• Building floor restriction
• Network latency
• Network bandwidth
• Physical distance

20
Minimum Cost Path
3

• Given multiple paths to a node, choose the
  shortest one

1
2
2
1
1
0
1
2
1
1
2
1
Note All links have a weight of 1
21
Limiting the Scope
3

• Define subtree of the network
• Bound the cost of each path from the reference
• Require strictly increasing path values

1
2
D2
2
1
1
0
1
2
1
1
2
1
Note All links have a weight of 1
22
Protocol Design Neighbor Discovery

• Beaconing mechanism determines near neighbors

Application
Network Abstractions
Neighbor Discovery
Monitor
23
Protocol Design Monitor Package

• Collect context information
• Local components to communicate directly with
  local sensors
• Remote components to communicate with sensors on
  other hosts
• Applications create monitors adhering to a
  recognized interface

Application
Network Abstractions
Neighbor Discovery
Monitor
24
Protocol Design Network Abstractions

• Network abstraction package implements the
  protocol, delivers queries, and returns responses
• Persistent and transient contexts are supported
• Responses return over either the tree of shortest
  paths or a mesh of acceptable paths

Application
Network Abstractions
Neighbor Discovery
Monitor
25
Egocentric Middleware View Concept

• Builds on network abstractions protocol for
  constructing contexts and distributing messages
• Provides abstraction of reference agents
  operating context, specific to its needs
• Contains a subset of the data items available on
  the hosts in the acquaintance list
• Uses constraints over profiles of hosts, agents,
  and data in addition to Network Abstractions

26
Conclusions and Future Work

• Declarative specifications provide flexible and
  general treatment of context
• Protocol shows feasibility of transparent context
  maintenance
• Weaker atomicity assumptions (e.g., w.r.t.
  configuration changes)
• Optimization that consider interactions among
  multiple contexts
• Formal verification
• Performance evaluation through simulations