Skip to main navigation menu Skip to main content Skip to site footer

Optimizing Backend Interactions in Microservices via Latency-Aware Design for Scalable Cloud Systems

Abstract

This paper presents a latency-aware backend interaction framework for microservices-based distributed systems. The framework combines asynchronous message queues, adaptive load balancing, and latency-aware scheduling to reduce blocking communication and coordinate resource allocation under dynamic workloads. Mathematical models describe communication cost, proportional workload distribution, and system throughput. Evaluation on a distributed microservices testbed compares the proposed framework with traditional synchronous and basic asynchronous queue-based methods. The reported results show an average latency of 70 ms, throughput of 1450 requests per second, and resource utilization of 85%, compared with 120 ms, 850 requests per second, and 65% for the synchronous baseline. Additional experiments examine workload variation, burst traffic, and partial node failures. The results indicate that coordinated communication, resource management, and scheduling can improve backend responsiveness, throughput, and resource efficiency in the evaluated environment.

Keywords

Microservices, backend interaction, adaptive load balancing, asynchronous communication, latency-aware scheduling, distributed systems

pdf