Your SlideShare is downloading. ×
Using trust-aware strategic agents for a self-organising computing grid
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×

Introducing the official SlideShare app

Stunning, full-screen experience for iPhone and Android

Text the download link to your phone

Standard text messaging rates apply

Using trust-aware strategic agents for a self-organising computing grid

262
views

Published on

Presentation by Yvonne Bernard at the 2nd Awareness Workshop on Challenges for Achieving Self-awareness in Autonomic Systems @ SASO 2012, Lyon, France

Presentation by Yvonne Bernard at the 2nd Awareness Workshop on Challenges for Achieving Self-awareness in Autonomic Systems @ SASO 2012, Lyon, France

Published in: Education, Technology

0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
262
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
4
Comments
0
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. Using trust-aware strategic agents for a self-organising computing grid Y. Bernard 10.09.2012 Awareness PhD Forum 1
  • 2. Outlineq  Motivation: Organic Compting system classq  Trust in OC systemsq  Application scenario: Trusted Desktop Gridq  Contributionq  Agent types and hierarchy [10] §  Static agents [3] §  Trust-adaptive Agents -  iTC Agents [5],[7] -  Evolutionary Agents [9] §  Strategic Agents [8]q  Adaptive Model of Observationq  Summary and OutlookY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 2
  • 3. Motivation: Organic Computing system class Example:q  New way of dealing with complexity Open Desktop Grid §  Self-X properties for decentralised solutions §  Incomplete system information Agent D §  Manage opennes -  Autonomous unknown agents Agent A X -  Selfish agents Agent C -  Malicious agents Agent Bq Implications to Trust facettes: §  Reliability/Functionality: Dynamic structure requires new approaches §  Security: Privacy and cooperation at the same time §  Safety: corrections during runtime possible §  Credibility: analyse environment at runtime §  Usability: Transparency and predictability not guaranteedà New class of algorithms necessaryY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 3
  • 4. Trustq  Trust := expectation value §  Probability that a certain event will happen in the future §  Reputation := Trust from indirect experienceq  Trust is a social mechanism, which allows more efficient and effective cooperation between individuals.q  This mechanism can be transferred into technical systems. §  Include trust aspect in cooperation decisionq  Trust as a constitutional part of technical systems §  Reduces information uncertainty in open systems (e.g. OC) §  Enables cooperation between subsystems (agents) -  increase efficiency of cooperating agents -  increase robustness regarding misbehaving agentsY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 4
  • 5. Application Open Desktop Grid Computing Agent A Agent G Agent B E F Agent F Agent C Agent E Agent D §  Computation on computers from different domains:Open system §  Free-riders refuse to accept work units. §  Egoists return wrong/incomplete results. §  Requires job replication and result checking -> inefficientY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 5
  • 6. Trusted Desktop Gridq  Decentralized system: All agents can §  offer computing resources (worker) and/or §  submit work units (submitter).q  Autonomous agents act on behalf of the users.q  Agents have a motivation to cheat.q  Basic Idea: enhance matchmaking with trust information §  Submitter: Who will be asked to process work units? §  Worker: Whose work units to accept?q  Goal: Enhance efficiency and robustness using trust and adaptation.Y. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 6
  • 7. Contributionq  Architecture for trust-adaptive strategic agents §  Generic model for OC and adaptive systems -  Adaptive to current situation -  Strategic decision making -  Model of Observation -  Institutional contol using constraintsq  Implementation of local trust-based adaptive strategy algorithms §  Submitter strategies §  Worker strategiesq  Evaluate architecture and matchmaking strategies and compare to Related Work (based on Grid metrics) §  H-Trust[12]: Trust- and Credibility-Tables §  Organic Grid[11]: Adaptive Tree OverlayY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 7
  • 8. Agent types High throughput/time Short makespan Decreased waste Decreased replication overhead Performance à Increasing observation overhead! Workload +Trust/Rep. +SD.S +SD.L AwarenessY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 8
  • 9. Agent types Pro-active trust-strategic agents Tactical agent eTC agent Adaptive MoO agent Reactive trust-adaptive agents Performance iTC agent Evolutionary agent Trust-aware agents Fixed stereotype agents Trust- neglecting agent Workload +Trust/Rep. +SD.S +SD.L AwarenessY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 9
  • 10. Agent hierarchyq  Static trust-considering agents: §  Behaviour prototypes: Free Rider, Egoistq  Trust-adaptive agents: reactive §  Adapt parameters to situation §  iTC Agent §  Evolutionary Agentq  Trust-strategic agents: proactive §  Tactical Agent: includes other agents expected behaviour §  eTC Agent: includes institutional control §  MoO Agent: Long-term strategic behaviour (access to predictions) -  Aim: find suited information/solution quality relation regarding overhead à adaptive Model of ObservationY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 10
  • 11. Adaptive Model of Observationq  Only evaluate information necessary for the current situationq  Overall goal: Reduce Overhead without sacrificing solution qualityq  Types of Overhead §  Communication: -  Update frequency (e.g. of reputation values) -  How many agents are asked to determine certain values (e.g. workload)? §  Calculation/storage: -  Aggregation of values -  Storing values for further evaluation (e.g. Time series analysis for prediction, relevant for strategic level)q  Adaptive cognition: select observed parameters based on §  role (submitter, worker) and §  situation (normal, increased attentiveness, alert)Y. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 11
  • 12. Summary and Outlookq  Trust can enhance communication, collaboration and negotiation in complex systems (e.g. OC systems)q  Application scenario Trusted Desktop Gridq  Approaches to trust-adaptive strategic agents [10] §  Static agents[3]: stereotypes of agent behaviour §  Trust-adaptive agents -  iTC Agents [5],[7] §  Efficient and robust §  Planned: Optimisation using learning techniques (thresholds) -  Evolutionary Agents [9]: first distributed learning approach §  Strategic Agents -  First approach: Tactical agent[8] -  Planned: eTC and MoO agent: §  Strategic Level on top of iTC Agents, institutional constraints §  Strategy based on long-term data and predictionY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 12
  • 13. Thank you for your attention!Y. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 13
  • 14. Publicationsq  [1] Martin Hoffmann, Michael Wittke, Yvonne Bernard, Ramin Soleymani, Jörg Hähner, DMCtrac: Distributed Multi Camera Tracking, ICDSC 08. Second ACM/IEEE International Conference on Distributed Smart Cameras, Sept. 2008.q  [2] Sven Tomforde, Martin Hoffmann, Yvonne Bernard, Lukas Klejnowski and Jörg Hähner, "POWEA: A System for Automated Network Protocol Parameter Optimisation Using Evolutionary Algorithms", Beiträge der 39. Jahrestagung der Gesellschaft für Informatik e.V. (GI), 2009, pp. 3177--3192, Gesellschaft für Informatik e.V. (GI)q  [3] Yvonne Bernard, Lukas Klejnowski, Jörg Hähner, Christian Müller-Schloer, "Towards Trust in Desktop Grid Systems", ccgrid, pp.637-642, 2010 10th IEEE/ACM International Conference on Cluster, Cloud and Grid Computing, 2010q  [4] Jan-Philipp Steghöfer, Rolf Kiefhaber, Karin Leichtenstern, Yvonne Bernard, Lukas Klejnowski, Wolfgang Reif, Theo Ungerer, Elisabeth André, Jörg Hähner, and Christian Müller-Schloer, "Trustworthy Organic Computing Systems: Challenges and Perspectives", Proceedings of the 7th International Conference on Autonomic and Trusted Computing (ATC 2010), Springerq  [5] Lukas Klejnowski, Yvonne Bernard, Jörg Hähner and Christian Müller-Schloer, "An architecture for trust-adaptive agents", Proceedings of the 2010 Fourth IEEE International Conference on Self- Adaptive and Self-Organizing Systems Workshop (SASOW 2010)Y. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 14
  • 15. Publicationsq  [6] Jan-Philipp Steghöfer, Florian Nafz, Wolfgang Reif, Yvonne Bernard, Lukas Klejnowski, Jörg Hähner and Christian Müller-Schloer, "Formal Analysis of Trusted Communities", Proceedings of the 2010 Fourth IEEE International Conference on Self-Adaptive and Self-Organizing Systems Workshop (SASOW 2010) q  [7] Yvonne Bernard, Lukas Klejnowski, Emre Cakar, Jörg Hähner and Christian Müller-Schloer, "Efficiency and robustness using Trusted Communities in a Trusted Desktop Grid", Proceedings of the 2011 Fifth IEEE International Conference on Self-Adaptive and Self-Organizing Systems Workshop (SASOW 2011) q  [8] Yvonne Bernard, Lukas Klejnowski, Ronald Becher, Markus Thimm, Jörg Hähner, Christian Müller- Schloer, "Grid agent cooperation strategies inspired by Game Theory", 4. Workshop Grid- Technologie für den Entwurf technischer Systeme, Dresden, 21.-22. September 2011, ISSN 1862-622X q  [9] Yvonne Bernard, Lukas Klejnowski, David Bluhm, Jörg Hähner and Christian Müller-Schloer, "An Evolutionary Approach to Grid Computing Agents", Proc. of the Italian Workshop on Artificial Life and Evolutionary Computation, 2012 , pp. 1-12, ISBN 978-88-903581-2-8 q  [10] Yvonne Bernard, Lukas Klejnowski, Jörg Hähner, and Christian Müller-Schloer, "Self-organising Trusted Communities of Trust-adaptive Agents", Awareness Magazine 2012, www.awareness-mag.eu, doi: 10.2417/3201011.004065 q  [11] A.J. Chakravarti, G. Baumgartner and M. Lauria. „The organic grid: self-organizing computation on a peer-to-peer network“. In: IEEE Transactions on Systems, Man and Cybernetics, Part A: Systems and Humans, 35.3 (Mai 2005), S. 373 –384. issn: 1083-4427. doi: 10.1109/TSMCA.2005.846396. q  [12] Huanyu Zhao and Xiaolin Li. „H-Trust: A Robust and Lightweight Group Reputation System for Peer-to-Peer Desktop Grid“. In: 28th International Conference on Distributed Computing SystemsY. Bernard Bernard Evolutionary Approach –240. Computing Agents Workshops. ICDCS10.09.2012 2008, S. 235 to Grid Y. An 08. Juni Awareness PhD Forum 15
  • 16. Outlookq  Controller §  Parameter optimisation (learning) on operational level §  Add long-term strategies (strategy) -  influence operational level §  Institutional control: Constraints -  Pre-filtering -  Post-filteringq  Observer §  Adaptive Model of Observation regarding: -  Which parameters are observed? -  Update frequency -  Agents sample size -  Memory size -  Aggregation method (time series analysis, Neural Networks,…)q  Compare trust-strategic agent with related work (H-Trust, Organic Grid)Y. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 16
  • 17. Trusted Manager O C Agent ConstraintsSD.L Strategic levelPredict(WL), Observer Controller WPredict(Trust) Long-term Situation SPredict(Rep) Strategic Decision Operational level Pre-selected Behaviour SD.S Observer Controller W WLTC, S TrustAgents, Current Situation Operational Decision RepAgents, RepOwn, Productive level Behaviour Worker Fitness Observer Controller Submitter Internal Situation Productive InteractionY. Bernard Bernard Evolutionary Approach to Grid Computing Agents Y. An 10.09.2012 Awareness PhD Forum 17