· Principles
· Blocking and stateful operators
· Streaming is batching and batching is streaming?
· Data windows
· Event time and processing time
· Exactly-once semantics
· Scaling to the enterprise level
Much of the discussion around batch versus stream processing centers on their high-level differences. However, the real distinction is more nuanced if we dig deeper. By closely examining the underlying principles of both data processing approaches, we can recognize their similarities. They actually have so much in common that it allows us to largely abstract away the technical differences.
This means that while application developers still need to choose whether their use case is better served by stream or batch processing style, they should not have to think much about the different technical implementation anymore. This would greatly simplify their primary focus on implementing business logic.
Be the first to comment