In the real world, applications are bound by rules such as data locality requirements, resource constraints, situations where a stable connection to the cloud cannot be guaranteed—such as at a baseball stadium, a construction site, or on a fighter jet—to do low-latency computation locally on the edge to avoid large amounts of data transfer. The application must be available and survive a regional disaster or cloud outage.
However, a need to run applications with the same consistency and stability that the Kubernetes platform provides still exists. Thus, solutions such as Anthos are designed to create this conformant distributed cloud platform for such applications.
In this chapter, we will go over these different situations and show various architectures involving the use of Anthos and its suite of products to support these types of applications.
This class of applications must run 24 hours a day, 7 days a week. Financial institutions managing transactions, health care applications, and traffic management are just some of the examples of applications that affect the real world if they are unavailable for a short period of time.