Data Streaming

Correct Answer: Data Streaming processes data in real-time, whereas batch processing processes data in large, finite batches.

The primary difference between Data Streaming and batch processing is the way data is processed. Data Streaming processes data in real-time, typically as soon as it is received, while batch processing processes data in large, finite batches, typically measured in hours or days. Data Streaming is well-suited for applications where low latency is critical, while batch processing is better for situations where data can be processed offline or in large, infrequent intervals.

Correct Answer: It enables the application to process data that arrives later than expected

Late-data handling mechanisms enable Data Streaming applications to process data that arrives after its expected arrival time. This is important for use cases where data can arrive with varying levels of latencies, and processing must be done within a certain time window. Late-data handling improves the accuracy of the application, and enables it to process additional data that may have been missed otherwise.

Correct Answer: A watermark is a way to specify the maximum timestamp that will be considered late in event time.

Watermarks are important when dealing with late data in event time processing. They allow us to specify a maximum timestamp that will be considered late, and any data that arrives later than that watermark will be discarded.