AWS Cloud Map

AWS Cloud Map is a cloud resource discovery service. Applications register resources such as databases, queues, object stores and microservices under custom names, and the service maintains their current locations as those resources change, continually checking their health so the recorded location stays accurate. An application then queries the registry for what it needs, filtered by attributes such as version and deployment environment.

The problem it addresses is stated plainly by the vendor and will be familiar to anyone who has maintained a growing system. Resource names and locations are usually managed by hand inside application code, which becomes time consuming and error prone as dependencies multiply or as services scale up and down with traffic. The alternative has been to install and operate a third-party discovery product, which is infrastructure a team then owns.

Registering arbitrary resources rather than only services is the difference from a service registry. A queue or a table is a dependency in exactly the way another microservice is, and being able to look both up through one mechanism removes a category of configuration.

Pricing is metered rather than subscribed, at a rate per registered resource per month plus a rate per million discovery calls, so cost tracks the size of the estate and how often it is queried.