Abstract
This paper examines service mesh technology as a dedicated infrastructure layer for managing interservice communication in microservice systems. The objective of the study is to analyze the architectural principles of network layer construction, estimate performance overhead, and demonstrate practical approaches for operationalization in Kubernetes clusters. It describes a model for separating the control plane and data plane, using the sidecar pattern for transparent traffic routing, implementing mTLS, and automatically collecting telemetry. Based on a literature review and open source documentation, a comparative analysis of three key implementations – Envoy, Istio, and Linkerd – is provided based on functionality, implementation complexity, and resource consumption. A practical example of an Istio configuration is provided, using Gateway, VirtualService, and DestinationRule manifests to manage HTTPS traffic and integrate with the Prometheus – Grafana – Jaeger – Kiali observability stack. It was shown that Service Mesh adds moderate overhead in terms of latency and resource consumption, but provides unified management of security and traffic between microservices. The obtained results allow us to formulate practical recommendations on the feasibility of using Service Mesh in enterprise microservice platforms, depending on the system scale and security requirements.