Analysis and Evaluation of Load Management Strategies in a Decentralized FaaS Environment: A Simulation-Based Framework

The Edge Computing paradigm has emerged to address new requirements in data processing. This approach enables the decentralization of computation by bringing computing capabilities to the edge of the network, i.e., close to the data sources, offering various benefits such as reduced latency and minimal network bandwidth consumption. In this context, Function as a Service (FaaS) emerges as a versatile and efficient solution, representing a specific instantiation of the Serverless Computing model. FaaS provides a scalable and reactive infrastructure that can be effectively applied to Edge Computing. Given the limited resource capacity of Edge nodes, appropriate load balancing is crucial. However, conducting on-field testing of any designed solution for this purpose can be arduous and time-consuming. This work addresses this challenge by proposing a simulation-based framework to design and evaluate load-management policies in decentralized FaaS environments. Additionally, we validate and compare four different load-balancing strategies, each characterized by varying degrees of complexity. Our experimental campaign demonstrates the effectiveness of the framework and our load-management methods across different operational scenarios.