A vital cloud computing paradigm that has the potential to greatly simplify application development while easing developers of the burden of operational responsibilities is serverless computing. That is further developed on serverless computing and that is known now as "Pay what is used" exist on to the clusters and require users to "Pay as you go." Researchers are paying huge attention to it as a representation of the progress of cloud programming models. The impact, migration of legacy applications, and adoption of this new methodology as industry standard are all addressed. In this paper, we evaluate the literature's description of serverless applications. We categorize applications into areas and talk in detail about the goals, practicality, and in each of those fields, the serverless paradigm presents challenges. and the serverless ecosystem's performance profile in a low latency, high availability setting, and give findings regarding the overall performance for improving serverless architectures. We focus our analysis just on AWS Lambda, one component of the serverless architecture used by AWS. Our findings demonstrate that Lambda-based architectures can have their performance characteristics tuned, and we discuss elements like cold starts and potential delay characteristics that may be brought about by the number of variables, including outside events and systems. In order to ultimately provide a collection of architectural principles that will enable serverless computing to be applied to a larger variety of problem areas, We suggest a broad range of tactics, ideas, and strategies that, when properly applied and deployed, simultaneously play to its advantages. We outline the areas that require additional study by the scientific community and identify challenging problems

cloud computing, serverless computing, legacy, applications, performance, AWS, Lambda, strategies, development, paradigm,challenging problems