1. Analyze challenges
Analyze and identify the challenges involved in solving the scalability problem of microservice-based applications
Objectives
General objective
The overall objective of this project is to overcome current limitations in elasticity management in microservice-based applications and to develop a generic elasticity controller that can be applied to a variety of applications without the need for retraining. To achieve this, we propose to design and develop an application latency predictor by combining Layered Queueing Network (LQN) and Graph Neural Network (GNN) techniques, as well as an elasticity controller that utilizes the developed predictor.
2. Analyze the state of the art
Gain a thorough understanding of the strengths and weaknesses of existing solutions regarding the elasticity of microservice-based applications
3. Predictor development
Design and develop a latency prediction model capable of effectively handling the highly dynamic and complex context of microservice-based applications.
4. Controller development
Design and develop an elasticity controller capable of reacting in real time to load changes and proposing efficient resource allocation.
5. Evaluation
Evaluate the proposed solutions across a series of selected scenarios and use cases to assess their effectiveness based on the identified requirements.