You have two choices on Amazon if you plan to run the services listed by you. Hortonworks Data Cloud for AWS is not one of them. It only applies to prescriptive use cases that include ETL, data discovery, data science and soon BI with Hive, Spark, Zeppelin and Druid, respectively. HDC also heavily relies on S3 for long term storage as opposed to hdfs. We did not include many options for the sake of simplicity. For your use case, you can look at Cloudbreak as my colleague noted as it has the similar experience to HDC with the benefit of infinite configurability as well as option to pass your own blueprints which can include the services you listed, including Storm and Kafka. CB relies on hdfs and is used for long running clusters. The other option you have is to spin up AMIs yourself and stand up HDP as you would an on premise cluster. Again both options give you full control and experience as on premise HDP but CB also offers you chance to use spot pricing, burst capacity to spin up or down nodes based on need, automation with CB shell.
