All the requirements we mentioned in the introduction to this chapter used to be coded within the application itself. As it turns out, many may be abstracted as they are shared across many different applications. When your application consists of many services, adding new features to all of them starts to be costly. With a service mesh, you may control these features from a single point instead.
Since a containerized workflow already abstracts some of the runtime and some networking, a service mesh takes the abstraction to another level. This way, the application within a container is only aware of what happens at the application level of the OSI networking model. The service mesh handles lower levels.
Setting up a service mesh allows you to control all network traffic in a new way and gives you better insights into this traffic. The dependencies become visible, as does the flow, shape, and amount of traffic.
Not only is the flow of traffic handled by the...
