SelfAdaptation for Everyday Systems Svein Hallsteinsen Erlend Stav

  • Slides: 10
Download presentation
Self-Adaptation for Everyday Systems Svein Hallsteinsen, Erlend Stav, Jacqueline Floch SINTEF ICT Norway ICT

Self-Adaptation for Everyday Systems Svein Hallsteinsen, Erlend Stav, Jacqueline Floch SINTEF ICT Norway ICT 1

Mobile use means changing context n When people are moving around using handheld networked

Mobile use means changing context n When people are moving around using handheld networked devices, the operating environment for the provided services vary n The user activity and location change, influencing user needs n The capabilities of computers and networks change, influencing the service quality properties n Applications depend upon system resources to satisfy user needs n When resources and needs vary, static applications fail to satisfy user needs most of the time n Changing context requires adaptive applications! influence user’s needs light noise position user context mobile user provided preferred mismatch! quality application affect operation system context battery shared devices network Qo. S ICT 2

Focus n Aspects of self-management we do address n Main: software architecture, mobile computing,

Focus n Aspects of self-management we do address n Main: software architecture, mobile computing, algorithms for system adaptation, adaptive components n Aspects we do not address n Prog. language support, AI techniques, autonomic computing, assurance n Targeted application types and properties n Mobile computing used in a changing environment n Everyday systems, for which non-perfect adaptation is acceptable n Cost effective development n Technical contributions n Overall architecture for model based adaptation middleware, combining a system family approach with context awareness and runtime adaptation n Uniform modelling of quality and context properties of components and composites ICT 3

Overall approach to adaptive applications describes relation light noise position influence user’s needs user

Overall approach to adaptive applications describes relation light noise position influence user’s needs user context mobile user architecture model selects application variant monitors preferred quality adapts adaptation middleware provided quality adaptable application components nodes monitors affect operation system context describes relation battery shared devices network Qo. S ICT 4

Overall architecture Application framework FAMOUS middleware Application instance FAMOUS adaptation middleware Architecture model 3.

Overall architecture Application framework FAMOUS middleware Application instance FAMOUS adaptation middleware Architecture model 3. consult Component repository Resource manager 2. plan Context monitor 1. change event Adaptation manager 4. plans Planner 6. change Meta model 7. change 5. configure Component instances Configurator FAMOUS component platform Distributed computing environment ICT 5

Architecture model Self-reliant client variant Medium client variant Client side o Server side Client

Architecture model Self-reliant client variant Medium client variant Client side o Server side Client side o Application or o o UI UI Medium client n o App n logic Server side n Self reliant client o App logic n and o Db n App. logic UI DB o Db n proxy Db proxy or UI (handsfree) o Db n UI (normal) App logic Db DB proxy Component repository ICT 6

Property annotation of variants Medium client variant haf=yes haf=no o o Hand c free

Property annotation of variants Medium client variant haf=yes haf=no o o Hand c free UI n c Normal UI n mem=30 mem=10 o c App logic o c n Db n avy = if n. nsb>80 then 100 else 100*(1 -(80 -n. nsb)/80) rsp = if n. nbw>80 then 10 else 10+100*(80 -n. nbw)/80) haf = UI. o. haf Client side Server side o mem=10 o UI Name Value range Explanation avy 1: 100 Availability of the service provided by the application rsp 1: 100 Response time mem 1: 100 Amount of memory nbw 1: 100 Bandwidth of network connection nsb 1: 100 Stability of network connection haf yes, no Hand free operation n o App n logic o Db n mem = UI. n. mem+Applogic. n. mem nbw =? nsb =? ICT 7

Utility function n Selection of variant is based on a utility function n User

Utility function n Selection of variant is based on a utility function n User preferences are to decide weight each properties c n o app n o exe usr utility = if n. mem > exe. o. mem then 0 else ( ( (if usr. n. avy <= app. o. avy then 1 else 1 - (usr. n. avy-app. o. avy)/usr. n. avy) + (if usr. n. rsp >= app. o. rsp then 1 else 1 - (app. o. rsp-usr. n. rsp)/app. o. rsp) + (If ((usr. n. haf and app. o. haf) or (!usr. n. haf and !app. o. haf) then 1 else 0 ) )/3 ) ICT 8

Remaining questions n How accurate do we need to model to achieve useful adaptations

Remaining questions n How accurate do we need to model to achieve useful adaptations for everyday systems? n Non-exact models acceptable as long as benefit are provided to the user n Is exhaustive search for best variant too expensive? n Resource constrained devices n Depends on number of variation points in application framework n Current approach focused on structural adaptations of components and connections within an application n How to extend the approach to multiple applications? n How to support other kinds of adaptation (e. g. adjustment of parameters with continuous range in components)? ICT 9

Organization and projects n SINTEF n Largest independent research organisation in Scandinavia with about

Organization and projects n SINTEF n Largest independent research organisation in Scandinavia with about 1700 employees n FAMOUS project (2003 -2006) n Strategic research project at SINTEF, funded by the Research Council of Norway n Includes a Ph. D work on context awareness n For more info: www. famous-project. net n MADAM project (2004 -2007) n EU-funded research project with industrial and university partners n Additional topics n n n Modelling tool support Network level middleware Industrial pilot applications and experiments ICT 10