Service mesh data plane: Any standard service mesh solution has to comply to this architecture of having two planes (control and data). The sidecar proxy is the data plane. Every request originating from any application service has to pass through this plane. The data plane is squarely responsible for performing service discovery, health checking, and routing. The data plane also does load balancing and authentication/authorization. Finally, the aspect of observability is indispensable for any service mesh solution. The data plane collects all kinds of performance, scalability, security, availability, and other decision-enabling information.

Service mesh control plane: The control plane monitors, configures, manages, and maintains all the data planes under its jurisdiction. The control plane elegantly provides policies and configuration for all the contributing data planes. The data planes form a kind of distributed system under the centralized monitoring of the control plane. That is the goal of distributed deployment, but the centralized monitoring is getting fulfilled through the leverage of this kind of architecture.