Cognitive Service Virtualisation: A New Machine Learning-Based Virtualisation to Generate Numeric Values.

Area of Science:

• Computer Science
• Software Engineering

Background:

• Continuous delivery (CD) requires readily available dependent services for effective quality assurance.
• Limited access to services and ripple effects from defects hinder CD adoption.
• Existing service virtualization (SV) approaches lack accuracy and ease-of-use for CD.

Purpose of the Study:

• To propose a novel machine learning-based approach for service virtualization.
• To enhance the accuracy of virtualised responses, specifically for numeric fields.
• To support the adoption of continuous delivery by improving service virtualization.

Main Methods:

• Developed a machine learning-based service virtualization approach (Cognitive SV).
• The approach predicts numeric field values using historical message data.
• Extended existing SV methods by incorporating machine learning for numeric data prediction.

Main Results:

• The Cognitive SV approach accurately predicts numeric fields in virtualised responses.
• Empirical evaluation demonstrated the approach's effectiveness across three datasets.
• The method successfully produced responses with appropriate fields, even for stateful protocols.

Conclusions:

• The proposed machine learning-based service virtualization approach effectively addresses CD challenges.
• Cognitive SV enhances the accuracy and usability of virtualised services.
• This contributes to more robust and efficient software development and deployment pipelines.