Citation Context

...or eliciting and specifying requirements about adaptation, the resulting artifacts do not explicitly refer to stateful goal instances that can be monitored from system traces. Different authors [26], =-=[27]-=- suggest using i* goal models to support the design of high-variability systems. The system chooses among existing options by evaluating the strategy that provides the best contribution to softgoals. ...

Citation Context

...d the associated situations can bessystematically identified and addressed via requirement analysissand elaboration, similar to identification of adaptation scenarios insa Dynamically Adaptive System =-=[17]-=-.sA challenge though is that due to the not fully anticipatedsenvironment, the adaptation cannot be fully determined at designstime. This calls for a solution, which in anticipated situationsswould pr...

Citation Context

...example, [6], [8] & [9]), it is more common for DASs to utilise some form of adaptive middleware (for example, [10] & [11]), thus separating the tasks of environmental monitoring, component substitution and state management from the system's business logic. Such adaptive middleware matches component selections with environmental conditions in accordance with adaptation policies, which specify the conditions under which components are substituted. Recently, researchers have investigated ways to understand and specify the requirements for a DAS, with a consensus emerging that goal-based methods [12] [13] offer a useful means to reason about how competing system goals are traded-off as the environment changes. There are several goal-based modelling approaches for DASs that utilise the popular KAOS [14] and i* [15] techniques ([12], [16] and [17] for example). However, despite the issue of DAS verification being highlighted as needing attention [3], there has been little work on the issue to date. The work that has been completed (e.g. [18]) has focussed on verifying business logic after adaptation, as opposed to our interest: verifying that the adaptation itself is desirable, or better st...

Citation Context

...rticular contexts, which features to select: when context choose features. A Goal-Based model will typically describe how features impact QoS properties, using for example help and hurt relationships =-=[6]-=-, and specify when QoS properties should be optimized (e.g., when a property is too low). Architecture Model: This model allows the designer to describe componentbased architectures. We can use any me...

Citation Context

...e an efficient runtime processing, but offer a limited scalability for dealing with the full specification of a complex adaptive system [3]. At a high-level of abstraction goaloptimization approaches =-=[4]-=-, [5], [6] propose to describe which (QoS) properties should be optimized in which contexts. The impact of each feature on these properties guides the selection of suitable configurations. These appro...

Citation Context

...e an efficient runtime processing, but offer a limited scalability for dealing with the full specification of a complex adaptive system [3]. At a high-level of abstraction goaloptimization approaches =-=[4]-=-, [5], [6] propose to describe which (QoS) properties should be optimized in which contexts. The impact of each feature on these properties guides the selection of suitable configurations. These appro...

Citation Context

...For requirements engineering, Berry et al. [3] have defined a framework of discourse for DAS requirements. Goal-based modeling has been used for specifying the adaptation choices that a DAS must make =-=[14, 22, 27, 43]-=- as well as for the specification of monitoring and switching between adaptive behaviors [31]. A particular strength of goal-based modeling is that it supports the modeling of nonfunctional trade-offs...

Citation Context

...ive systems (e.g., [47, 31]). In addition, some research approaches have successfully used goal models as a foundation for specifying the autonomic behaviour [29] and requirements of adaptive systems =-=[22]-=-. One of the main challenges that self-adaptation poses is that when designing a self-adaptive system, we cannot assume that all adaptations are known in advance — that is, we cannot anticipate requir...

Citation Context

...lection. Fickas and Feather [23] provide a framework for systems to monitor their executions and modify themselves at runtime to better satisfy stakeholders’ goals. Similar approaches are proposed in =-=[21, 30, 25]-=-. Several frameworks have been developed for the generation of software monitors from requirements models [21, 35, 33, 17, 37]. Specifically, [37] goes a step further allowing not just monitoring but ...

Citation Context

...omprises solutions for modeling the requirements of service-based, and adaptive, systems, tracing them, and overseeing the reliability of compositions. As for the first aspect, some preliminary works =-=[5]-=- have already used goal models for specifying the requirements of adaptive systems. Their adaptation strategies are expressed as enumerations of predefined alternative tasks. Our solution, in contrast...