Amazon Web Services released its newest service to the public today: a data stream collecting hundreds of terabytes of data per hour and processing it in real time.
Amazon Kinesis gives developers a range of data capture and analysis tools to manipulate massive quantities of data in real time. In a scenario where developers are tasked with monitoring application-performance logs, responding to minute-by-minute financial data or identifying trends in a social media stream, Kinesis automatically manages infrastructure, storage, networking and deployment to enable instantaneous data analysis.
Kinesis also comes with delivery and storage integration options to other Amazon services such as S3, Redshift and DynamoDB, along with elastic scalability for various-sized data streams and a GitHub-hosted client library for creating Kinesis applications.
As an example of real-time analytics capabilities, take an online shopping application. During high-volume shopper loads, developers can use Kinesis to quickly identify click-through paths, customer behavior or slow queries to drive recommendations and generate alerts. If log data streaming into Kinesis indicates errors from a user, it could relay those logs to a backup system, or send information on to a relevant developer.
Kinesis also integrates with Amazon CloudWatch for monitoring, displaying metrics like throughput, utilization and latency in the AWS Management Console.
Developer resources detailing how to use the various features and control data operations in Amazon Kinesis include the Developer Guide and API reference in addition to the open-source client library.
Kinesis is currently only available in the “US East, North Virginia region,” but where available the stream data is stored for 24 hours, during which it can be read, reread, backfilled, analyzed, or moved to long-term storage. Kinesis is priced at approximately 15 cents per hourly shard or throughput rate, in addition to existing node usage fees.