Optimizing Networking

We have data persistent with storage in the cloud. We also have compute resources to perform tasks on that data. Now, we need some form of plumbing to connect these components if they are to communicate with each other and move from one place to another. Networking is the plumbing that will serve these functions and is the topic at hand in this chapter. More specifically, we’ll look at ways to optimize our networking workloads on AWS.

The underlying system architecture will incur different networking costs. Distributed systems, for example, which span multiple availability zones (AZs) and multiple Regions for redundancy, will have higher networking costs than a system that resides entirely in one AZ. It’s hard to apply blanket network optimization strategies compared to mechanisms such as savings plans, which can potentially cover all compute usage within an AWS organization. We’ll discuss ways to incorporate networking cost considerations...