How does the "Embedded Security" model differ from the classical cryptographic model?

Question

Accepted Answer

In the embedded security model, the attacker can attack everywhere — not just the communication channel, but also the endpoints themselves.

Aspect
Classical Model
Embedded Security Model

Attack surface
Communication channel only
Channel AND endpoints

Crypto operations in
Black box (tamper-proof HW)
Grey box (e.g., protected software)

Security depends on
Algorithm strength + key size
Algorithm + key size + protocol choice + secure implementation

The evolution of the trust boundary:

Old model: Only the channel between Alice and Bob is untrusted
Current model: The attacker is everywhere — channels, endpoints, even inside devices

This is why secure implementation matters as much as strong algorithms. A mathematically perfect cipher can be broken through:

Poor implementation
Side-channel attacks
Protocol weaknesses

Aspect	Classical Model	Embedded Security Model
Attack surface	Communication channel only	Channel AND endpoints
Crypto operations in	Black box (tamper-proof HW)	Grey box (e.g., protected software)
Security depends on	Algorithm strength + key size	Algorithm + key size + protocol choice + secure implementation