Analyzing runtime adaptability of collaboration patterns. (28th November 2014)

• Record Type: Journal Article
• Authors: Dorn, Christoph; Taylor, Richard N.; Smari, Waleed W.; Navarro, Antonio M.; McQuay, William K.
• Is Part Of: Concurrency and computation. Volume 27:Number 11(2015:Aug.)
• Page Start: 2725
• Page End: 2750
• Publication Date: 2014-11-28
• Subjects: Parallel processing (Electronic computers) -- Periodicals; Parallel computers -- Periodicals; 004.35
• DOI: 10.1002/cpe.3438
• Languages: English
• ISSNs: 1532-0626

Abstract

<abstract abstract-type="main" id="cpe3438-abs-0001"> <title>Summary</title> <p id="cpe3438-para-0001">The recent two decades have witnessed the emergence of large‐scale, interaction‐intensive systems. A system's provided user‐centric communication and coordination mechanisms have a significant impact on its runtime management. Beyond a certain size, manual monitoring and management are no longer feasible. Hence, it is highly important for a system designer to becoming aware of the most suitable interaction mechanisms and their implications on system adaptability. Specifically, a system designer requires knowledge on what adaptation primitives are available, whether these are system‐driven or user‐driven, how long they will take, what impact do they have on collaboration state, and under what conditions they can be enacted. These aspects vary considerably across collaboration patterns.</p> <p id="cpe3438-para-0002">In this paper, we investigate a collaboration structure's adaptability based on behavior, asynchrony, state, and execution context. We subsequently discuss seven distinctively different collaboration patterns in terms of those aspects. Based on a motivating scenario, we ultimately demonstrate how these patterns and insights into their inherent adaptability may guide design decision impact and trade‐off analysis. Copyright © 2014 John Wiley &amp; Sons, Ltd.</p> </abstract>