Explain how Amazon Kinesis works for real‑time data streaming and how you can run Spark code on Kinesis streams.

Amazon Kinesis is a fully managed service that lets you ingest, buffer, and process streaming data in real time. A Kinesis Data Stream is a set of shards; each shard can ingest up to 1 MB/s of data and read up to 2 MB/s. Producers write records to the stream, and consumers read them in parallel. To run Spark on Kinesis, you use Spark Structured Streaming with the Kinesis connector. The connector reads records from the stream as a streaming DataFrame, applies transformations, and writes results to a sink such as S3, Redshift, or another stream. The typical workflow is: 1) Create a Kinesis stream and configure a producer (e.g., Kinesis Agent, SDK). 2) In Spark, use spark.readStream.format("kinesis") with the stream name, region, and credentials. 3) Apply Spark SQL or DataFrame APIs to process the data. 4) Write the output using writeStream. Spark handles checkpointing and fault tolerance automatically. This integration allows you to build real‑time analytics pipelines that scale horizontally with the number of shards and Spark executors.