Linkerd has a dedicated layer 7 proxy that deals with HTTP and HTTP/2 for requests and responses. It can use a filter chain for these requests for success, failure, latency, and responses. A service mesh implies that you deploy one Linkerd proxy alongside a microservice. When you initiate a service call, instead of it being direct, that request is received by the Linkerd proxy and then sent to the microservice. Next, the microservice response is routed through the Linkerd proxy, which again sends that response to another microservice. The proxy sitting next to each microservice wraps the network call and collects the metrics. Service-to-service communication is secured through TLS, and all the traffic on the wire is encrypted. 

Linkerd provides an abstract layer so you can manage, control, and monitor microservices. Linkerd facilitates a service-oriented infrastructure through load balancers, TLS, request routing, and service scalability to make applications resilient. Linkerd Proxy allows application owners to develop microservices in the programming language of their choice.

There are two primary architecture components to Linkerd for deploying and running standalone proxies:

• Control plane 
• Data plane 

Linkerd's control and data plane architecture can be seen in the following diagram:

In this chapter, we will understand this architecture. Let's begin with the control plane.