Cloud & platform

Service mesh

An infrastructure layer of proxies that handles service-to-service traffic, giving mTLS, identity, authorisation and telemetry without application changes.

A service mesh puts a proxy beside each workload (a sidecar such as Envoy, or a per-node proxy in ambient designs) and a control plane that configures them. Every call between services goes proxy to proxy, so the mesh can enforce mTLS with short-lived workload certificates (often SPIFFE IDs), apply authorisation policy by caller identity, and emit uniform metrics and traces.

For security it is the cheapest way to get encrypted, authenticated east-west traffic and identity-based policy at scale. It turns "which services may call payments" into config instead of code in every team.

Where it goes wrong: permissive mode left on forever (plaintext still accepted), policy that authenticates but never authorises, and assuming the mesh covers traffic that bypasses the proxy (host-network pods, direct egress). It adds latency and an operationally critical control plane. In interviews, explain what the mesh proves (workload identity) and what it does not (end-user authorisation).

Learn it properly

Also used in

Related terms