The most powerful cloud computing service model currently being developed is the Serverless computing model. This model brings scalability and cost optimization in the process of deploying applications on cloud infrastructure. The application will be divided into functions that run a specific logic and those functions will be deployed as independent units on the Serverless computing platform. One of the strongest points of deploying applications running on the Serverless computing platform is its scalability and good processing performance. Scalability is demonstrated through the process of recognizing the actual usage needs of the function, from which the controllers in the Serverless computing platform will calculate and coordinate resources in the cloud environment appropriately, ensuring both service availability and saving idle resources. The performance of functions deployed on the Serverless computing platform mainly comes from the computing and storage capacity provided by the infrastructure, but the Serverless platforms also participate in supporting the process of optimizing the operating flow to minimize processing time and return response results to requests from the function caller. In this paper, we will build an experimental model with the two most popular open source Serverless computing platforms in the cloud computing development community, OpenFaaS and Knative. The purpose of this work is to compare and evaluate the scalability and performance in the process of operating applications on the open source Serverless computing platform. These two platforms rely on two different parameters to decide on the number of function instances.
