2012/07/25

Cellular tissue autonomic behavior



Cellular Autonomic Behavior computing paradigm:

A new approach for agent based autonomic computing, in complement of use the autonomic nervous system, use a cells tissue as the conceptual foundation. In this concept exists one physical component type called cell and a logical component called tissue, multiple cell forms a tissue and have proprietary knowledge of its state, the tissue is a composition of cells and have the knowledge of the whole tissue that no one single cell have. The cells provides services.

A set of modifications on the Axiomonitor agent based software to embed autonomic features:

Cell level functions:

  • Cell level self healing 
    • over computing detection
    • ask for reinforcements
    • service restart
  • Cell level self configuring
    • by internal demand
    • by tissue demand
    • cell based services 
    • effectors and sensor
  • Cell level self protecting
    • intrusion detection
    • resource allocation limits
    • service isolation schemes
  • Cell level self environment knowledge
    • CPU
    •  Network
    • Disk
    • Memory
    • Usable external resource
  • Cell communications
    • with tissue (nervous system)
    • direct with other cells (data osmosis)
  • Cell level service deployment 
    • requisites validation
    • Code deployment


Tissue level:

  • Tissue level resources availability
    • service high availability
    • service contingency
  • Tissue level service migration
    • movement to other infrastructure
    • tissue reconstruction 
      •  lost (dead) cells
    • leave failing or exiting cells
  • Tissue level service/resource allocation on demand
    • by cell reinforcement asking
    •  by average response time  increase
  • Tissue level service fail over
    • redundancy of services provided by the tissue
  • Tissue level self configuration 
    • record usage patters and provide enough resource to meet the performance goal.
    • use service allocation
    • use service migration
  • Tissue level service demand prediction
    • records usage patterns and processing times
    • by expected performance definition establishes a relation between allocated resource and response times
  • Tissue level external change management integration
  • Tissue monitoring activity expose
    • Exposes activities to humans
    • Permits changes to decisions of reconfiguration


Inga kommentarer:

Skicka en kommentar