Chapter 11. Serving layer: Illustration

Having covered the requirements of the serving layer, we can now consider an example of a database built specifically for use as a serving layer database. As with all the illustration chapters, no new theory is introduced in this chapter; it instead serves to map the concepts you’ve learned to the nuts and bolts of real tooling.

We’ve mentioned that the available tooling for the serving layer lags behind the ideal possibilities, and that fact will be evident as you build the serving layer for SuperWebAnalytics.com. You’ll learn to use ElephantDB, a key/value serving layer database. Because it doesn’t support index types beyond key/value, you’ll have to diverge from the ideal index types described in the previous chapter.

We’ll examine the basic architecture of ElephantDB to understand how it meets the requirements of the serving layer, and then we’ll review its API to retrieve the contents of a batch view. Finally, you’ll see how you can use ElephantDB to index and serve the batch views for SuperWebAnalytics.com.

11.1. Basics of ElephantDB

ElephantDB is a key/value database where both keys and values are stored as byte arrays. ElephantDB partitions the batch views over a fixed number of shards, and each ElephantDB server is responsible for some subset of those shards.

11.2. Building the serving layer for SuperWebAnalytics.com

11.3. Summary