Using a single PUT API call to Firehose, companies can get an entire stream of data from devices into AWS's Redshift database service and/or Simple Storage Service (S3). Once the data hits Firehose, Amazon will send it to the selected storage service or services within a minute.
Firehose is designed to elastically scale with the data that users put in, which means that companies don't have to worry about provisioning more instances to keep up as they add devices.
Once the data is inside S3 or Redshift, it can be connected to other products that Amazon has, including the company's new QuickSight business intelligence service, which was also announced Wednesday.
While working with the Internet of Things is a key and core functionality to the product, it's also capable of pulling in real-time metrics on digital marketing and advertising campaigns, and can enable mobile applications to push data directly into AWS where it can then be pulled into analytics tools and dashboards.
Companies are charged for their use of the service based on the size of each data record they send through Firehose, rounded up to the nearest 5KB. For example, if each data record is 53KB, they'll be billed as though they were sending 55KB records. Amazon then charges US$0.035 to $0.038 per gigabyte ingested through the service, depending on what region users are sending the data to.
Amazon already offered similar functionality using its Kinesis Streams service, but actually connecting IoT devices and applications to that service requires that companies build a custom application. That takes a lot of time, and that's time businesses may not have or want to devote to a project like one that Kinesis Firehose is useful for.
It's also something that will help Amazon compete with Microsoft Azure's IoT Suite, a bundled group of services that are designed to help companies get data from embedded sources into that cloud and then process that data inside Azure.