Windowing (Streaming Analytics)

What it is: Grouping streaming events into time-based “windows” for calculations.

What it’s for:

• Compute metrics like “average speed per 1 minute” or “count events per 5 minutes”.
• Make continuous streams measurable in time slices.

Common window examples:

• Tumbling window (non-overlapping): every 1 minute
• Sliding window (overlapping): last 5 minutes updated every 1 minute

Exam cues:

• “compute metric per minute / per time interval” → windowing.

Hard words:

• *window* /ˈwɪndoʊ/: cửa sổ thời gian
• *tumbling* /ˈtʌmblɪŋ/: rời rạc (không chồng lấn)
• *sliding* /ˈslaɪdɪŋ/: trượt (có chồng lấn)
• *interval* /ˈɪntərvl/: khoảng thời gian