101 Challenges for Self-awareness in Autonomic Systems

1. How can distributed systems with no central controller become collectively self-aware, rather than at individual node level? 2. To understand self-awareness in autonomic systems we must first understand the boundaries of self-over time, context and scale 3. Incentivizing users to cooperate by providing access to location data/social groups to study natural human mobility 4. Create collective embodied systems where self-healing emerges in response to adverse internal/external conditions 5. In systems with dynamic service composition, how can we achieve system-level self-awareness of service components? 6. Sensors, sensors, sensors: given the volume of interesting data available, how can services understand what they need 7. Evolving a collective system that exhibits self-awareness and environmental awareness from scratch 8. The role of conservation laws in collective awareness-exchange of mass & energy vs. exchange of information 9. Evolving the step from environmental awareness to self-awareness 10. Find construction rules of artificial self-aware systems by revealing the common core in natural collective systems 11. Grand challenge in Self-Awareness? Real-world apps, with real hard requirements - best research driver there is! 12. To build better models to understand the basic principles of self-* 13. To build a better theory to analyze the data from the models and from real world 14. To develop methodologies and tools to engineer systems 15. To define when autonomic systems are beneficial or detrimental to a given domain or application 16. To bring computers near a level where humans are, not humans down 17. To model, test and verify self-aware systems 18. To develop techniques to control self-organization 19. How we can learn from human self-awareness? 20. How to make aware components behave to reach a global optimum? 21. To combine computer science with social science 22. To be aware of neighbors 23. To be aware of what awareness actually means 24. To analyze the emerging patterns in evolving behaviors 25. How do components make themselves aware of the surrounding (open) environment? 26. Systems that exhibit self-awareness as emerging properties 27. Collective self-awareness from not self-aware components 28. How can we say that a system is self-aware? 29. To make systems actually know what happens inside them 30. How to engineer the system to produce the correct emergent behavior? 31. Definition and metrics of self-awareness 32. To define real grounded application scenarios 33. To model the context and to validate the model itself 34. What actually is a self-organizing system and how to build it? 35. To have an operation definition of self-awareness 36. Adaptability, evolvability, diversity, spatiality 37. To develop ubiquitous platforms 38. To bring together experimental and theoretical communities 39. How modeling can be considered in the development? 40. To have efficient computation 41. How is this research going to contribute to the challenges of global warming and sustainability? 42. Using competition across the fields to push the research further and faster 43. How robot controllers (mind) and mechanical parts (body) can co-evolve? 44. How do we formally understand what trust is? 45. How can services understand what they really need? 46. How to know whether a system is aware and the issue of global and local awareness?

Monitoring and controlling emergent properties and specifying and controlling adaptation How to improve the communication between local and global systems in swarm robotics? The difficulty of proving all the properties of an emergent system The difficulty of writing precise requirements about flexibility Measuring and finding metrics for the different kinds of awareness How to measure the level of awareness? E.g. The number of variables AND the algorithm that processes the information from the variables? 53. Letting different systems interoperate and collaborate 54. Considering sociological aspects besides technical aspects 55. Dealing with high and low levels of contexts 56. How to map raw data to knowledge? 57. How to engineer decision systems? 58. How to achieve adatpivity at runtime? 59. How to manage the relationship between individual and group levels? 60. How to ensure safety and correctness? 61. How to describe and to compare information? 62. To enable adaptive systems to learn online 63. To have an appropriate mathematical base 64. To exploit a graphical language in order to achieve automatic generation of engines 65. To have good and sustainable test bed and test environment for experiments 66. Introducing economic models 67. Monitoring of large scale adaptive infrastructures and mobile devices 68. To disambiguate the awareness concepts 69. Checking, requirements, model, verification and validation at runtime 70. Representation and synchronization of requirements at runtime 71. To address real problems by means of exemplars 72. To have intelligent runtime environments that support adaptation, keeping and managing the model also at runtime 73. Controlling the system behavior by its model 74. How to measure adaptiveness? 75. How awareness emerges in complex systems? 76. Understanding people activities and needs 77. How to improve safety by means of awareness? 78. How can self-aware systems enhance the human collective intelligence? 79. How to implement self-awareness in artificial systems? 80. How to measure self-awareness? 81. Awareness of peers in P2P 82. Joining energy-awareness and security 83. Energy-awareness: how to gather the energy information from all system components in real time 84. Which applications benefit from self-awareness, and how are they improved by it? 85. What makes a system or application self-aware, and what are essential components? 86. Can we identify different levels or types of self-awareness? 87. What is the definition of self-awareness in a technical or engineering context? 88. Not defining if a system is self-aware or not, but in which aspect(s) exhibits self-awareness 89. Connection between self-awareness and environment-awareness 90. Which are the metrics and the benchmarks to define the level of self-awareness? 91. How can we guarantee that the systems fulfill the self-awareness specifications? 92. We must demonstrate that self-awareness is a useful concept, showing what we cannot do with state-of-theart methods 93. How to grant correct behavior in the worst case? 94. How to exploit self-awareness for users' satisfaction? 95. How to exploit self-awareness in modeling complex systems? 96. We should measure self-awareness on the base of users' expectations 97. The metrics of awareness must consider also the environment of the measured system 98. Can we distinguish between platform self-awareness and application self-awareness? 99. To select the useful data for self-awareness 100. How to handle the huge amount of information and choices? 101. Convince humans that self-aware systems are not against them

47. 48. 49. 50. 51. 52.

Contributors to the 101 Challenges: Marco D. Santambrogio, Jason Miller, Alessandro Nacci, Tom Davidson,
Lamia Youseff, Christian Plessl, Dirk Koch, Jeong-A Lee, Marco Platzner, Christian Plessl, Stefan Wilderman, Lamia Youseff, Marco D. Santambrogio, Tomoya Enokido, Mauro Migliardi, Juan Manuel Orduna, Alan Winfield, Jose Halloy, Ichiro Satoh, Simon Dobson, Mirko Viroli, Giovanna Di Marzo Serugendo, Sara Montagna, Carlo Ghezzi, Luciano Baresi, Nelly Bencomo, Hausi A. Mller, Ramana Reddy, Ivona Brandic, Nenad Stojnic, Tom Keeley, Franco Bagnoli, Peter Lewis, Yvonne Bernard, Manuele Brambilla, Carlo Pinciroli, Martin Wirsing, Henry Bensler, Emil Vassev, Wei Dai, Francois Toutain, Guillame Dugue, Gusz Eiben, Franco Zambonelli, Jose Luis Fernandez, Matthias Holzl , Martin Wirsing , Rocco De Nicola, Gabriella Castelli , Alois Ferscha, Evert Haasdijk, Julie McCann, Jeremy Pitt, Frederic Gruau, Alan Brook, Colette Johnen, Stefan Dulman, Akla-Esso Tchao, Giuseppe Valetto, Ingo Scholtes, Daniel Dubois, Marco Mamei, Paul Snyder, Christopher Hollander, Rolf Kiefhaber, Xinghui Zhao, Andres Ramirez, Jean Botev, Venkatraman Iyer, Frank Schweitzer, Julia Shaumeier, Nils Rosemann, Holger Prothmann, Glen Fink, Cortney Riggs, Sven Bruckner, Tom Holvoet, Thomas Schmickl, Giacomo Cabri, Jon Timmis, Walter Colombo, Ben Paechter, Emma Hart

At SASO 2013 Awareness hosted an activity concerning the 101 Awareness Challenges. The result of this activity (below) details what participants viewed as the most relevant challenges and suggestions for possible research directions to face them:

Challenge

1. How can distributed systems with no central controller become collectively selfaware, rather than at individual node level? 30. How to engineer the system to produce the correct emergent behavior?

Research direction

Methods to validate that the self-awareness is a system property

Challenge

21. 54.

To combine computer science with social science Considering sociological aspects besides technical aspects

Research directions

Funding interdisciplinary projects Use of test-beds and experiments Research in big data and social data Languages or design tools for awareness-oriented programming

Challenge

53. 59.

Letting different systems interoperate and collaborate How to manage the relationship between individual and group levels?

Research direction

Research on engineering systems

Challenge Research directions

71. To address real problems by means of exemplars Understanding the principles of the systems Interdisciplinary Involvement of the industry

Challenge Research direction

74.

How to measure adaptiveness?

Definition of models, metrics and tools to measure the degree of adaptiveness of systems