A service mesh is an application communication layer that standardizesservice-to-service trafficinside Kubernetes. Instead of each development team building custom logic for retries, timeouts, encryption, and telemetry, the mesh provides these capabilities consistently across workloads. This is typically done by inserting a data plane (often sidecar proxies or node-level proxies) that intercepts inbound and outbound traffic for each microservice, plus a control plane that distributes configuration and identity material.

The key outcomes align directly to optionB: communication becomespossible(reliable connectivity patterns),structured(consistent routing rules, policies, and identity), andobservable(metrics, logs, and distributed tracing for east-west traffic). A service mesh commonly adds controls such asmTLS encryption, fine-grainedtraffic policy(allow/deny, rate limits, circuit breaking), and progressive delivery patterns (canary/blue-green) without changing application code.

By contrast, service discovery (A) is usually a built-in Kubernetes function, load balancing/autoscaling across sites (C) is not the primary definition of a service mesh, and a single centralized global routing table (D) is not how meshes are typically described or implemented.