Other sources for input to Storm

In our earlier example, we have seen integrated with Storm, one of them being Kafka (just discussed in the previous section). In Storm samples, the word count topology (which we covered in detail in an earlier chapter) doesn't use any data source for input. Instead, some sentences are hardcoded in the spout itself and this seed data is emitted to the topology. This may be fine for test cases and samples, but for real-world implementations this is neither ideal nor expected. Storm has to feed a stream of live events into the topology in almost all of the real-world implementations. We can have a variety of input sources that can be integrated with Storm. Let's have a closer look at code snippets on what all we can plug in with Storm to feed the data.

A file as an input source

We can use a Storm spout to be effectively reading from a file; though that's not a real use case for a streaming app, but we can have Storm very well read in from the file. All we need...