State Management in Serverless Workflows using Durable Functions on Azure

Abstract

This paper examines the integration of Azure Durable Functions with advanced state management techniques to address the challenges faced by serverless workflows, particularly in event-driven architectures. Serverless computing simplifies infrastructure management but struggles with maintaining state across distributed function executions. To overcome this, the paper proposes a hybrid state management framework that combines stateless function execution with externalized state stores, such as Redis and DynamoDB. The framework aims to ensure fault tolerance, scalability, and low-latency processing in long-running workflows. Through experimental validation, the paper demonstrates the potential of this framework in enhancing the efficiency and reliability of cloud-native applications. The study also explores the practical implications of this solution for industries such as healthcare and finance, highlighting its importance in the context of modern event-driven and cloud computing systems.