Serverless Optimisation Tool

Reference number 10841

Sectors: Software & IT Services

Industries: Software

Technology overview

Building upon the R&D work of a large EU consortium, RADON is a DevOps framework to create and manage microservices-based applications that can optimally exploit serverless function-as-a-service computing technologies. RADON applications leverage the industry-standard OASIS TOSCA to abstract the cloud and edge deployment characteristics, allowing the user to create and manage the lifecycle of complex microservices-based architectures. A distinguishing feature of the framework is its use of model-based abstractions, which make it easy to re-engineer applications for migration between different deployment environments, as well as to support diverse customer needs and integration scenarios. Compared to other industrial serverless frameworks, RADON offers a comprehensive toolchain that covers several unique assets:

  1. a repository of pre-defined TOSCA templates to accelerate deployment on OpenFaaS, Amazon AWS Lambda, Azure Function, Google Cloud Platform, Docker, Raspberry PIs.
  2. CDL, a logic-based language to express deployment requirements, with an associated verification tool to verify compliance;
  3. a deployment optimization feature to select the right concurrency and memory assigned to functions, to minimize deployment costs;
  4. a feature to transform TOSCA models of monolithic software systems into corresponding, highly-decomposed, FaaS-based products.

The above tools leverage RADON TOSCA, a custom-built extension of OASIS TOSCA to enable serverless function deployment and orchestration. The framework comes with documentation for each tool and many tutorial examples.


Lulai Zhu, Giorgos Giotis, Vasilis Tountopoulos and Giuliano Casale.

RDOF: Deployment Optimization for Function as a Service. IEEE CLOUD 2021.

G. Casale et al., RADON: rational decomposition and orchestration for serverless computing, Springer SICS, 35, pages 77–87 (2020).


Dr Giuliano Casale

Faculty of Engineering, Department of Computing

Visit personal site

Contact us about this technology


James Chan

Industry Partnerships and Commercialisation Senior Executive

James joined Imperial in Sep 2021 as an Industry Partnerships and Commercialisation Senior Executive. He primarily takes care of IP and licensing cases from Department of Electrical and Electronic Engineering and Department of Computing. Prior to joining Imperial, James worked in The University of Hong Kong and The Chinese University of Hong Kong for more than a decade handling around several hundreds of inventions and a few thousands of patent applications/patents from the Faculties of Engineering, Science and Medicine, including a patent portfolio with multiple concurrent exclusive licensees, which generated more than 80% of the licensing income of the university, in the magnitude of mid-10 millions of US dollars per year. He also helped the university to develop the licensing policy for copyright materials, includes, software, database, evaluation/diagnosing questionnaire, multimedia files, etc. With his extensive experience in Tech Transfer, James was also invited to teach a module on “Intellectual Property Management and Technology Commercialisation” under the MSc in Electronic Engineering programme and part of the module on “Technology, Society and Engineering Practice” of the undergraduate programme of CUHK. James is an Electronic Engineer by training, with degree from CUHK and Bachelor of Laws specialised in IP Laws from The University […]

Contact James

[email protected]

Related technologies

MOFsoft: software modelling for optimal MOF fabrication

MOFsoft: software modelling for optimal MOF fabrication

Our model solves an important inverse design problem in Microstructured Optical Fibre (MOF) fabrication: how to find the initial preform shape and the manufacturing parameters required for the desired final cross-plane geometry. Find out more

Serverless Optimisation Tool

Serverless Optimisation Tool

The serverless optimization tool helps to reduce the cost of operating serverless software systems. Leveraging a model-based approach, the tool first predicts the performance of a serverless-based application. Then, using numerical optimization it recommends the optimal configuration of each function such that the total operating cost is minimized while meeting target performance requirements. Find out more

Sign up for updates

Sign up for monthly technology alerts via email, and find other ways to connect with us.