Big Data

Chapter 10. Serving layer

This chapter covers

• Tailoring batch views to the queries they serve
• A new answer to the data-normalization versus denormalization debate
• Advantages of batch-writable, random-read, and no random-write databases
• Contrasting a Lambda Architecture solution with a fully incremental solution

At this point you’ve learned how to precompute arbitrary views of any dataset by making use of batch computation. For the views to be useful, you must be able to access their contents with low latency, and as shown in figure 10.1, this is the role of the serving layer. The serving layer indexes the views and provides interfaces so that the precomputed data can be quickly queried.

Figure 10.1. In the Lambda Architecture, the serving ...