A DSL toolkit for deferring architectural decisions in DSL-based software design

Lecture Notes in Computer Science

Navarro

Leymann

2017

Self Cite

Microservice-based software architecture design has been widely discussed, and best practices have been published as architecture design patterns. However, conformance to those patterns is hard to ensure and assess automatically, leading to problems such as architectural drift and erosion, especially in the context of continued software evolution or large-scale microservice systems. In addition, not much in the component and connector architecture models is specific (only) to the microservices approach, whereas other aspects really specific to that approach, such as independent deployment of microservices, are usually modeled in other views or not at all. We suggest a set of constraints to check and metrics to assess architecture conformance to microservice patterns. In comparison to expert judgment derived from the patterns, a subset of these constraints and metrics shows a good relative performance and potential for automation.

Section: Discussionmentioning

confidence: 99%

Ensuring and Assessing Architecture Conformance to Microservice Decomposition Patterns

Lecture Notes in Computer Science

Navarro

Leymann

2017

Self Cite

“…We implemented a prototype model-driven process monitoring framework to demonstrate our DSL. The prototype is built on top of an MDD framework (Frag [25]). …”

Section: Compliance Monitoring Domain-specific Language (Dsl)mentioning

confidence: 99%

Domain-specific language for event-based compliance monitoring in process-driven SOAs

Mulo

Dustdar

2012

SOCA

Self Cite

Organizations today are required to adhere to a number of compliance concerns from laws, regulations and policies. Compliance is achieved through defining and implementing so-called controls in the organizations' business processes. Organizations that build their systems based on the process-driven SOA paradigm realize business processes through orchestration of services to handle the process' business activities. These business activities or groups of business activities in some cases realize the compliance controls. We propose an approach for implementing event-based compliance monitoring infrastructure that observes such business processes to verify that compliance is indeed adhered to. Our approach is essentially a model-driven technique for realizing this infrastructure. We implement a domainspecific language for specification of compliance directives, and we include code generation templates to generate compliance monitoring code, which is leveraged by complex event processing components to monitor for compliance. We evaluate the impact of our approach on the effort and productivity of a developer who is specifying compliance directives.

“…In terms of technical implementations, our event view model and the DSL are realized using Frag [19]. Frag is a dynamic programming language designed, among other things, to build (meta-)models and DSLs.…”

Section: Approach Overviewmentioning

confidence: 99%

“…The specification of this and other event view model instances is done using the eventing DSL. As the eventing DSL is embedded into the Frag language, any extensions to the event view model can be quickly reflected in the eventing DSL through the modeling framework from Frag [19]. Hence, we would not require extra effort to extend tools like parsers to understand the DSL extensions.…”

Section: Service Create Watchmeservice\mentioning

confidence: 99%

An event view model and DSL for engineering an event-based SOA monitoring infrastructure

Mulo

Proceedings of the Fourth ACM International Conference on Distributed Event-Based Systems

Dustdar

2010

Self Cite

An event-based solution that uses events to convey information to a monitoring tool is well suited to implementing a non-intrusive monitoring infrastructure. This enables an SOA system's stakeholders to observe the system aspects of interest to them. However, implementation of SOA today, let alone the monitoring infrastructure, is a complex task due to the heterogeneous environment consisting of multiple technologies, platforms and components. We propose an approach for implementing such an event-based SOA monitoring infrastructure, that introduces a dedicated event view model and an eventing domain-specific language in a model-driven framework. The event view model captures SOA artifacts and links them with the event domain, while the eventing domain-specific language enables a system developer to specify instances of the event view model. With our model-driven approach, most of the runtime monitoring infrastructure is generated. These two ingredients (view model and domain-specific language) focus implementation efforts on the concern of eventing, thereby helping to manage complexity. We apply and evaluate our approach in the context of a case study.