Visitor from Vienna University of Technology Explores Software Architecture Styles and Collaboration Patters

In March 2011, Christoph Dorn from the Vienna University of Technology (TU Wien) in Austria joined ISR as a post-doctoral visiting researcher.  Dorn completed his Ph.D. on Adaptation Techniques in large-scale Service-oriented Systems in 2009 and continued as a post-doctoral researcher at the Distributed Systems Group (TU Wien). Christoph Dorn  His general research focus is on adaptation and recommendation mechanisms in socio-technical systems.  In such systems, Humans are no longer just the ‘users’ of a system but also an integral part.  Their interactions with other humans and software elements have a significant impact on the runtime management of software components.  Dorn’s recent work addressed heuristics for team formation in social networks, self-adaptive people-driven workflows, and interaction-driven service composition.

Dorn choose UC Irvine/ISR as the host institution of his Erwin Schroedinger Mobility grant because of its reputation in ground breaking research and innovation in the domain of software engineering, especially software architectures for large-scale, adaptive systems.  Software architecture styles have a profound effect on system adaptability, a research area that was fundamentally shaped by faculty host ISR Director Richard N. Taylor.  A principled, architecture-drive approach is thus of utmost importance when addressing the complexity of socio-technical systems.

During his stay Dorn will focus on analyzing the interdependencies of software architecture styles and human collaboration patterns, and modeling such relations during system design and system run-time. The ultimate goal is deriving adaptation mechanisms for socio-technical systems that are aware of the human collaboration structure when reconfiguring the supporting software system.

Mapping between software architecture styles and collaboration patternsTo this end Dorn has collaborated with Prof. Taylor in the recent months to outline a mapping between software architecture styles and collaboration patterns (see table, right).  The result is a technical report detailing the similarities of styles such as publish-subscribe, components and connectors (C2), or Representational State Transfer (REST) with collaboration patterns such as Master-Worker, Organizational Control, or Self-organizing Teams.  The report also provides a detailed analysis of two large-scale systems: Amazon Mechanical Turk (Map-Reduce) and Twitter (Publish-Subscribe).

Preliminary insights from this technical report highlight the potential synergies between architecture styles and collaboration patterns.  Adaptation mechanisms addressing human collaborations are candidates for designing new techniques for managing software systems comprising massive, autonomous, unreliable components.  Individual, self-governing components cannot be directly manipulated; they accept only recommendations.  They might follow the recommendations or choose to ignore them based on internal, unobservable constraints.  Subsequently, human-inspired properties such as trust, reputation, or cooperativity potentially become applicable to software components.  Also, mimicking human trust establishment  through joint, repeated interactions provides an alternative, self-regulating security approach in open software systems.

CREST team members Strasser, Baquero, and GorlickThe technical report, entitled “Mapping Software Architecture Styles and Collaboration Patterns for Engineering Adaptive Mixed Systems,” can be downloaded from the ISR website (UCI-ISR-11-4,

Dorn’s upcoming research efforts will focus on integrating architecture description languages (ADL) and collaboration pattern models.  While exploring design of adaptation mechanisms, he plans to collaborate with ISR Ph.D. students Michael Gorlick, Kyle Strasser, and Alegria Baquero (R. Taylor, advisor) who are currently working on the CREST framework.

Christoph Dorn is visiting ISR through May 2012. He can be reached at

His publications and additional background information are available at:

This article appeared in ISR Connector issue: