Raw data are read and separated by system code (aka “parsers”) into data streams (science, engineering, metadata, etc.) based on content. Each stream contains multiple parameters, including a time stamp. A single instrument can produce several data streams with many different parameters within them.