Embedded system types and definition difficulty

An embedded system is a computer system (processor, memory, I/O) built into a larger product to perform a specific dedicated function, usually with real-time and resource constraints.

Types

Based on performance and functional requirements:

• Real-time embedded systems: hard real-time (missing a deadline is a failure, e.g. airbag, pacemaker) and soft real-time (occasional miss tolerated, e.g. video player).
• Stand-alone embedded systems: work on their own, e.g. microwave oven, MP3 player.
• Networked embedded systems: connected to a LAN/WAN/Internet, e.g. ATM, IoT devices, home security.
• Mobile embedded systems: portable, e.g. mobile phone, digital camera.

Based on processor size/complexity: small-scale (8/16-bit microcontroller), medium-scale (16/32-bit) and sophisticated (32/64-bit with RTOS, configurable processors).

Why it is hard to define

• It is found in so many forms (a toy, a car, a satellite) that no single description fits all.
• Many embedded systems are now general-purpose and programmable (a smartphone), blurring the line with a PC.
• Embedded computers are hidden inside other products, so it is unclear where the "computer" ends and the device begins.
• Size, cost, performance and software complexity range from a few bytes of code to millions of lines.
• A PC can also be used as an embedded component, so it is easier to define by exclusion ("any computing system that is not a desktop or laptop") than by features.