1 / 34

Ralf König LMU München, Germany May 15 2009

Dagstuhl Seminar „Self-healing and self-adaptive systems“. Engineering of IT management automation along tasks, loops, function allocation, implementation method catalog. Ralf König LMU München, Germany May 15 2009.

rea
Download Presentation

Ralf König LMU München, Germany May 15 2009

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Dagstuhl Seminar „Self-healing and self-adaptive systems“ Engineering of IT management automation along tasks, loops, function allocation, implementation method catalog Ralf König LMU München, Germany May 15 2009

  2. How can we replace IT mgmt automation as an art with a more step-wise, engineering-style method along steps? Idea: Combine knowledge from: • IT management • Systems engineering • Automation engineering • Engineering in general

  3. Applications for system design

  4. Vision: IT mgmt autom. workbench Task analysis and task allocation • Mgmt automation building blocks • loops and loop steps • machine cap’s • (human cap’s) Requirements engineering Managed IT resources Design constraints Cost benefit analysis Model forecasts Simulation of behavior

  5. Q1: What are relevant entities in IT mgmt automation and how do they relate to each other?

  6. The essence of IT mgmt automation Tasks People Implementation methods IT resources

  7. Split of concept and implementation Resources x Tasks Implementation methods

  8. Q2: Can we give tasks a structure?

  9. Step 1: Task Analysis Sources of inspiration: (Albus, NIST: A Reference Model Arch. for Intelligent Systems Design, 1970’s) (Sheridan, Humans and Automation, 2002)

  10. Tasks and task decomposition • Hierarchical task decomposition • tree structure • rooted in sys. engineering and robotics • proposed by Sheridan (MIT) and Albus (NIST) • Workflow based task decomposition • E.g. in BPMN • Fits well with IT mgmt! • BUT: No common internal structure!

  11. Q3: Is there a common structure, that we can align all tasks to, that groups sub-tasks according to functionality?

  12. Step 2: Common loops Sources of inspiration: Sheridan, Kephart

  13. 2/3: Loop alignment - Use MAPDEK!

  14. 3/3: Also Control Loops fit well

  15. Q4: How to handle task allocation tohumans/machines? How to design levels of automation?

  16. Step 3: Allocation to Human/Machine Inspiration: Sheridan

  17. Levels of Automation Source: Sheridan, Humans and Automation, Wiley, 2002.

  18. Q5: How to categorize the implementation methods?

  19. Step 4: Implementation Method Catalog • Decide • N-modular redundancy • voting schemes

  20. Impl. method catalog by behavior (loops) • monitor - monitor some managed object (often an event queue or set of configuration parameters), display the events/parameters in some view • configurator (often called wizard) - group parameters into logical groups, provide configuration interface at a higher level of abstraction • solution adviser according to customer-level requirements • notifier monitor some managed object, send message when applicable • watchdog monitor some managed object, trigger alarm or predefined action when applicable • load balancer balance incoming load by assigning parts of the load to resources in a pool • filter monitor an event queue, drop certain events • stabilizer protect some system parameter from external distortions • optimizer optimize some system parameter regarding some goal function • classifier put incoming events into one of several categories • planner plan the next steps to reach some goal, build a plan • scheduler take jobs, assign start and end times in a schedule • simulation environment

  21. Conclusion Task analysis and task allocation • Mgmt automation building blocks • loops and loop steps • machine cap’s • (human cap’s) Requirements engineering Managed IT resources Design constraints Future work: Use Matlab/Simulink for an update mgmt automation example Cost benefit analysis Model forecasts Simulation of behavior

  22. Thank you for your attention! Ralf König <koenig@mnm-team.org>

  23. Paper titles, example 1

  24. Conclusion: Eng. IT mgmt automation • Transfer concepts from systems engineering of classical automation domains with to IT management automation! • Input of the method: • IT management automation scenario • Tasks • Operator Roles • Resources • Policies and Design Constraints • Output of the method: • Task decomposition into loops • Loop decomposition into loop steps along MAPDEK struct. • Selection from a catalog for the allocation of loop steps to humans/machines • Selection from a catalog of methods to implement machine-executed loop steps • Out of my (!) scope: • Non-resource related IT management tasks • Economic effects of automation (e.g. cost reduction) • Staffing (choosing the right people for the right job) • Software implementation of the individual machine functions 1. 2. 3. 4. 5.

  25. Request by organizers • Avoid conference-style talks! • These tend to provide more answers than questions. • Instead, present questions, show your current ideas and insights, but not all the details. • These are better discussed personally by the few people who are really interested!

  26. Paper titles, example 2

  27. Paper titles, example 3

  28. Paper titles, example 4

  29. Q1: What are relevant entities in IT management automation?

  30. Entities in IT mgmt automation

  31. The main entities

  32. Loops 1/3: Basis: MAPEK from ACI Source: An architectural blueprint for autonomic computing. 4th ed., IBM, June 2006.

  33. A large solution table is a bad idea

  34. IT Mgmt and Engineering fit together!

More Related