Optimized strategy in cloud-native environment for inter-service communication in microservices

Abstract

Microservices architecture has become the most popular architecture for building scalable and flexible applications due to its inherent advantages over monolithic and Service-Oriented Architecture (SOA) systems. It allows developers to scale individual components independently, improve maintainability through smaller codebases, and increase resilience by isolating service failures. Additionally, microservices are well- suited for cloud-native environments, supporting containerization and orchestration platforms. However, transforming monolithic systems into microservice-based systems introduces challenges with inter-service communication, leading to performance bottlenecks and increased latency due to the decentralized and distributed nature of the architecture. This research addresses the above challenge by proposing a novel communication model for cloud-native environments, focusing on optimizing inter-service communication in microservice architecture. The proposed strategy operates on top of the TCP network layer rather than the application layer, leveraging TCP streams but using a request-response-based communication model. All distributed systems typically rely on TCP-based communication over a network. To reduce the overhead of opening and closing socket connections, the proposed strategy persists the TCP connection when a microservice starts and maintains it until the service goes down. This solution reduces network resource consumption by serializing messages into binary form, thereby enhancing response times and application throughput. When sending messages over the TCP connection, we have established a method for ensuring exact-once delivery, improving the reliability and efficiency of message transfer. The proposed solution was implemented within a microservice architecture and deployed in a cloud-native environment to assess its effectiveness for cloud-native applications. The solution has been evaluated against the traditional HTTP-based communication method in the same microservice-architected environment and has shown improved performance in terms of latency and overall throughput. This research offers a robust communication model for microservices that addresses performance challenges while maintaining compatibility with cloud-native concepts. The proposed approach is a step toward optimizing microservices communication, enabling developers to build scalable, efficient, and resilient applications in distributed environments.