Streaming real-time data has many use cases such as log analytics and IOT analytics. In most of these cases, however, the throughput requirement for the producer may be high and actual consumption rate might require only a lower throughput. We could use a single stream with a high throughput, but then we will also be wasting throughput. Therefore, we can have different streams with different throughputs, as required, and then, using Kinesis Analytics' capabilities (for example, analyzing data streams with SQL), we can route the request into an appropriate stream.