Summary

In this chapter, we have talked about the batch model serving pattern. This is the first pattern that we have categorized based on the serving approaches we discussed in Chapter 2, Introducing Model Serving Patterns. We introduced what batch serving is and looked at the different batch serving patterns based on the batch job triggering strategy and the inference strategy. We explained, with examples, how cron expressions can be used to schedule batch jobs. We discussed some examples where batch serving is essential. Then, we looked at an end-to-end example of batch serving using a dummy model, persistence store, and server.

In the next chapter, we will discuss online serving, which updates the models immediately with new data.