Serverless computing has become a prominent approach for developing cloud-based applications. In this paradigm, developers create small, stateless functions designed to handle specific events within the cloud infrastructure. The Function-as-a-Service (FaaS) model of serverless computing enables applications to be structured as independent, fine-grained units known as functions. This approach allows developers to concentrate solely on application development while delegating resource management responsibilities to cloud service providers. The widespread adoption of serverless computing is largely driven by its key features, including automatic scalability, on-demand resource allocation, event-driven operation, and a cost-efficient pay-as-you-go billing system. This study focuses on the heightened latency experienced in serverless computing environments, arising from the diverse challenges inherent in the architecture and implementation of serverless systems. A major drawback of serverless computing lies in its lack of emphasis on data. Current serverless frameworks predominantly rely on a data shipping architecture. This chapter presents an alternate architecture to state-of-the-art data shipping architecture of serverless computing. The proposed architecture enables the computation to be performed at the side of data by permitting the code to flow across regions. Another limitation linked to serverless computing is the “limited lifespan” of containers. Therefore, managing incoming function requests results in an increased number of container creations, leading to cold starts. An extensive solution to handle the added up latency due to the cold start problem is proposed in our work. As the second contribution of this chapter, it introduces a scheduling strategy aimed at minimizing cold start events by extending the lifespan of containers. This is achieved through the Least Recently Used Warm Container Selection (LCS) method combined with scheduling based on affinity-driven worker node selection. Further, an orchestration service is crucial for running applications in a serverless environment. We suggest an orchestration service with the ability to access resources from different regions simply by deploying the application in a single region. Additionally, this orchestration service optimizes data transmission latency without requiring deployment in multiple regions. The orchestration service not only gives access to inter-region resources but also optimizes the performance of application by restricting the communication latency.