Choosing a location
A connected device can send information to a remote computing system, process it locally, or divide the work. Edge computing describes approaches that bring processing closer to the data source, such as a device or nearby computer.
The distance is not simply geographical. A nearby processor may reduce the communication needed before a result becomes available. Sending selected information onward can also reduce the amount of raw data crossing a network.
Edge and central computing can work together, dividing immediate calculations from a wider stored history.
A useful division of work
Imagine a hypothetical room monitor that reads temperature throughout the day. A nearby processor could summarize the readings, then send a periodic summary to another system.
The design still needs to account for limited local resources, updates and the information available at each location. Moving a calculation closer does not automatically make every part of a system faster, safer or independent of a connection.
Ask what needs an immediate answer, what can wait, and what information has to move between locations. These questions make the arrangement easier to understand than the word “edge” alone. The central issue is the relationship between a task, its data and the place where it runs.
Further reading: IBM: what edge computing means.