...

1. ETL - An development team has some realtime Hydrator pipelines that use a CDAP Worker. They want to run their ETL pipelines using Spark Streaming because their company is standardizing on Spark.

2. Data enrichment - Every time a purchase is made on an online store, an event with purchase information is pushed to Kafka. The event contains a timestamp, purchase id, customer id, item id, and price. A pipeline developer wants to create a realtime pipeline that reads events from Kafka and joins customer information (email, age, gender, etc) to each event, then writes the events to a CDAP Table.

3. Window and Reduce - Customer locations are Apache log data is being fed into Kafka. Each event contains the customer id and their locationresponse code, response time, url, and other fields. A pipeline developer wants to create a pipeline that reads these events and transforms it into records containing the customer id, speed of the customer, distance of the customer from their home, a flag indicating whether or not the customer just left their home, and demographic information about the customer (age, gender, etc).writes a set of output records every 10 seconds, where each record contains a url, and the min, max, and average response times for that url over the last five minutes.  
4. Machine Learning - An email client is set up to push an event to a Kafka topic whenever somebody uses the client to send an email. The client is also set up to push an event to another topic whenever an email is marked as spam. A pipeline developer wants to create a realtime pipeline that reads events from spam topic and trains a spam classification model in realtime using Streaming linear regression (http://spark.apache.org/docs/latest/mllib-linear-methods.html#streaming-linear-regression). The pipeline developer also wants to create another pipeline that reads from the email topic and adds a 'isSpam' field to each record based on the model trained by the other pipeline.
5. Ordered Event Processing - Customer locations are being fed into Kafka. Each event contains the customer id and their location. A pipeline developer wants to create a pipeline that reads these events and transforms it into records containing the customer id, speed of the customer, distance of the customer from their home, a flag indicating whether or not the customer just left their home, and demographic information about the customer (age, gender, etc).  Note: this use case doesn't seem possible without doing some weird things in SparkStreaming

User Stories

1. As a pipeline developer, I want to create realtime ETL pipelines that run using Spark Streaming. (3.5)

2. As a pipeline developer, I want to be able to group events into time windows in my streaming pipeline. (3.5)

3. As a pipeline developer, I want to perform aggregations over windows of events in my pipeline. (3.5)

4. As a pipeline developer, I want to enrich streaming events by joining to other datasets. (3.5)

5. As a pipeline developer, I want to join data streams in my pipeline. (3.5)

6. As a pipeline developer, I want to train machine learning models in my streaming pipeline. (4.0)

7. As a plugin developer, I want to be able to create new streaming source and sink plugins. (3.5)

8. As a plugin developer, I want my transform, aggregator, joiner, and sparkcompute plugins to work in both Spark Streaming and Data Pipelines. (3.5)

9. As a plugin developer, I want to be able to write plugins that train ML models using MLLib. (4.0)
   
   

Design

We will introduce a new artifact similar to the DataPipeline artifact, called the DataStreaming artifact.  It will use the exact same configuration, except it will support its own set of plugin types. This artifact will support the transform, sparkcompute, aggregator, and joiner plugin types.  In addition, we will add streamingsource, streamingsink, and streamingtransform plugin types. 

...