Abstract
Emerging serverless computing technologies, such as function as a service (FaaS), enable developers to virtualize the internal logic of an application, simplifying the management of cloud-native services and allowing cost savings through billing and scaling at the level of individual functions. Serverless computing is therefore rapidly shifting the attention of software vendors to the challenge of developing cloud applications deployable on FaaS platforms. In this vision paper, we present the research agenda of the RADON project ( http://radon-h2020.eu ), which aims to develop a model-driven DevOps framework for creating and managing applications based on serverless computing. RADON applications will consist of fine-grained and independent microservices that can efficiently and optimally exploit FaaS and container technologies. Our methodology strives to tackle complexity in designing such applications, including the solution of optimal decomposition, the reuse of serverless functions as well as the abstraction and actuation of event processing chains, while avoiding cloud vendor lock-in through models.
Introduction
Serverless computing, particularly Function-as-a-Service (FaaS), offers significant benefits in terms of operational efficiency and cost-effectiveness by abstracting server management and providing granular scaling. However, these benefits come at the cost of increased architectural complexity. Developers face significant hurdles in decomposing monolithic applications into hundreds of event-driven functions, managing vendor-specific APIs that lead to lock-in, and ensuring quality in highly distributed, ephemeral environments. The RADON project addresses these challenges by proposing a holistic, model-driven DevOps framework that unifies serverless functions, microservices, and data pipelines.
The RADON Architecture
The RADON framework is structured across several layers to support the entire lifecycle of a serverless application, from requirement definition to runtime management.
As shown in the architecture diagram, the framework spans four primary domains:
- Requirements: Utilizing a Constraint Definition Language (CDL) to specify functional and non-functional goals, such as performance and cost.
- Design & Code: Providing a graphical modeling environment and an IDE (based on Eclipse Che) for defining system topologies and Infrastructure-as-Code (IaC) recipes.
- Runtime Environment: Managing orchestration, function repositories, and monitoring feedback loops to ensure continuous delivery.
- Cloud: Abstracting the underlying infrastructure, whether it be serverless FaaS, microservices, or data pipelines, to prevent vendor lock-in.
Model-Driven DevOps and Rational Decomposition
At the core of RADON is the extension of the OASIS TOSCA (Topology and Orchestration Specification for Cloud Applications) standard. While standard TOSCA focuses on static infrastructure topology, RADON introduces behavioral modeling to capture the temporal sequences and event-processing chains inherent in FaaS.
"Rational decomposition" refers to the use of formal models to determine the optimal granularity of services. Instead of arbitrary splitting, architects can use the framework to analyze how different decomposition strategies impact performance, cost, and security. This is supported by an abstraction layer—event gateways—that decouples event producers from consumers, allowing the same model to be deployed across different cloud providers.
Automated Quality Assurance
RADON integrates quality assurance directly into the DevOps pipeline, treating Infrastructure-as-Code (IaC) with the same rigor as application code. Key tools include:
- Defect Prediction: Employs machine learning to identify "bad smells" and potential bugs within IaC scripts, trained on extensive open-source codebases.
- Load Testing: Automatically generates realistic workloads based on production monitoring data and input traces to verify Service Level Agreements (SLAs).
- Decomposition Optimization: Uses logic programming and simulation to find the best way to aggregate or split services to meet specific performance targets.
Industrial Use Cases
The framework's utility is demonstrated through three diverse industrial scenarios:
- ADAMO (Tourism Promotion): Migrating legacy Java backends to serverless to handle highly variable tourist traffic from mobile applications.
- SARA (Ambient Assisted Living): Processing IoT data from wearables for elderly care, requiring low-latency, event-driven responses to health events.
- Orbit (Managed DevOps): Automating complex CI/CD pipelines for external clients to reduce manual intervention.
Conclusion and Future Outlook
RADON provides a roadmap for managing the complexity of serverless architectures through rigorous modeling and automated tooling. While challenges such as state space explosion in formal verification and the mitigation of "cold starts" remain, the project establishes a foundation for vendor-agnostic, high-quality serverless DevOps. Future work focuses on refining the Constraint Definition Language and maturing the open-source toolset for broader community adoption.
Summary updated