Streams

A stream is a channel or flow through which material moves. In the natural environment, a stream is a flowing body of water contained within a bed and banks, typically a tributary of a river. In computing, the term is used to describe a sequence of data elements that become available over time, allowing programs to process elements as they arrive rather than after a fixed data set has been collected.

Natural streams vary in size and flow, are part of hydrological networks, and influence ecosystems and landforms. They are measured by discharge and flow rate and are affected by precipitation, geology, and land use. Stream order and watershed boundaries describe their position in drainage systems and help assess factors such as erosion, sediment transport, and nutrient cycling.

In computing, streams provide a uniform interface for reading or writing data. They can be classified by origin (input versus output), by data unit (byte streams for binary data, character streams for text), or by buffering and encoding. Typical environments expose standard streams such as standard input, output, and error, and many languages offer abstractions for network, file, and memory streams. This abstraction enables consistent data handling across different sources and destinations.

Data streams are potentially unbounded sequences of data elements produced over time by sources such as sensors, log files, or message queues. Stream processing aims to compute over elements as they arrive, supporting stateless and stateful operations, windowed aggregations, and low-latency responses. This contrasts with batch processing, which operates on finite datasets. Streaming systems often rely on message brokers or distributed processing engines and address challenges such as out-of-order events and fault tolerance.

The concept also extends to streaming media and other continuous data-delivery contexts, where the emphasis is on timely, progressive data consumption. Across domains, streams are characterized by temporal sequencing, ongoing data flow, and the need for incremental, often real-time, processing.