Microservice (MS) architecture has become a popular paradigm in software engineering and has been embraced in the industry (e.g., Amazon, Netflix) for cloud-based applications with crucial performance requirements. Surprisingly, assessing how the MS designs affect performance is still a challenging issue, which is generally tackled by extensive and expensive profiling. In this paper, we propose mu P, a novel development framework for MS applications where performance can be predicted by design. mu P offers an API that automatically generates a performance model based on Layered Queuing Networks (LQNs) without requiring any development effort beyond writing the actual system code. The model can then be queried to predict performance metrics such as response time and utilization of individual microservices. We validate mu P on four benchmarks taken from the literature. The results show the effectiveness of mu P in accurately predicting performance due to increasing user load, vertical and horizontal scaling. We report prediction errors for response times consistently lower than 10% across a wide range of operating conditions.
μP: A Development Framework for Predicting Performance of Microservices by Design
Garbi, G;
2023-01-01
Abstract
Microservice (MS) architecture has become a popular paradigm in software engineering and has been embraced in the industry (e.g., Amazon, Netflix) for cloud-based applications with crucial performance requirements. Surprisingly, assessing how the MS designs affect performance is still a challenging issue, which is generally tackled by extensive and expensive profiling. In this paper, we propose mu P, a novel development framework for MS applications where performance can be predicted by design. mu P offers an API that automatically generates a performance model based on Layered Queuing Networks (LQNs) without requiring any development effort beyond writing the actual system code. The model can then be queried to predict performance metrics such as response time and utilization of individual microservices. We validate mu P on four benchmarks taken from the literature. The results show the effectiveness of mu P in accurately predicting performance due to increasing user load, vertical and horizontal scaling. We report prediction errors for response times consistently lower than 10% across a wide range of operating conditions.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
