Mobile Adaptive Tasks Guided by Resource Contracts

1
Mobile Adaptive Tasks Guided by Resource Contracts

• Peter Rigole
• K.U.Leuven Belgium

2
Limited resources

• Ubiquitous computing environments realized by
• Mobile devices
• Small devices hidden in paintwork
• Occasionally supported by powerful systems
• Mainly resource-limited devices
• Tradeoff
• More functionality
• More resources in use
• Self-managing software in an environment with
  limited resources requires adaptive behavior

3
Resource aware Components

• Component-based applications
• fine-grained applications enhance
  self-manageability
• delineated entity for code analysis
• Resource-Awareness per component

Resource Delaration
Resource Contract Proposal
Resource Delaration
Resource Delaration
Agreement between Several parties e.g. two
component ports
Negotiated and validated contract proposal gets
signed
4
Dynamic Task Models

• There is no seamless course of events in our
  lives
• Seams are at the origin of the unpredictable
  nature of life
• Taking real-world events into account
• By interpreting context information
• gt Evolution of a task according to the users
  context

• Taking users agenda into account
• gt Scheduling of tasks

5
Towards a Deployment Plan (DP)

• A DP specifies
• which software components have to be loaded and
• how they should be interconnected
• A DP should
• follow the evolutions in the task schedule
• guiding internal adaptations
• take resource availability into account
• the DP is influenced by external adaptations
• relocation, load balancing
• work in a continuously changing environment
• resource availability changes
• tasks change
• gt Task models are a starting point for composing
  a DP
• gt Resource information must be considered !

6
Task Model to Deployment
foraging constraint solver
7
External Adaptations

• Middleware interferes with the application
  without guidance of the application itself
• Is transparent for the application
• Types
• Component reconfiguration
• changing configuration parameters
• Component replacement
• new version
• Component relocation
• Foraging

8
Internal adaptations

• Application changes itself
• mainly driven by context information and task
  models
• knowledge about the evolution of the users task
• User task changes, thus functionality changes,
  thus application can change as well
• removal of unnecessary components
• adding new component for new functionality
• adding/removing of optional components

9
Ubiquitous adaptation

• Coordination between
• Internal adaptations
• External adaptations
• Guided by contracts to ensure the validity and
  integrity of
• Application compositions
• local and
• distributed

10
General Idea

• Realizing the idea of Calm Computing by
• Adding a dynamic task model into the pervasive
  comp. env.
• Contains info about actions to undertake
• Mapping the model onto a deployment plan
  automatically
• Guiding an internal adaptation mechanism
• Evolution without user distraction
• Using a foraging technique (relocation)
• Expanding an applications computing space autom.
• Is an external adaptation mechanism
• Task driven deployment and foraging are driven by
  
• resource information (as context information)
• negotiating resource contracts
• to do the adaptations automatically

11

• Questions ?

Peter.Rigole_at_cs.kuleuven.ac.be