Is a Three-Layer Post-Quantum Safe VPN Hidden Within Regular IPsec Effective Against Eavesdroppers?

Question

I'm working on designing a VPN that is post-quantum safe while avoiding detection that it uses post-quantum cryptography. The goal is to make the use of post-quantum cryptography indistinguishable to an eavesdropper, who might otherwise store the traffic for future decryption attempts using quantum algorithms once they become available. An attacker might be especially interested in post-quantum encrypted traffic because, when someone puts in the effort to use post-quantum cryptography at all (currently), the data transmitted can be assumed to be extremely sensitive.

Here’s the approach I'm considering:

Layer 1: Use regular IPsec encryption for the initial layer, as is common in current VPN implementations.

Layer 2: Within the IPsec tunnel, encrypt the data using a post-quantum encryption algorithm.

Layer 3: Re-encrypt the post-quantum encrypted data with another layer of regular IPsec. The idea is that the outer IPsec layer makes the traffic look like standard encrypted traffic. If an eavesdropper intercepts this, they would only see regular IPsec-encrypted data, not realizing that within it, there's a layer of post-quantum encryption.

Here are the reasons for this approach:

Obfuscation: By wrapping post-quantum encrypted content within regular IPsec, the use of post-quantum cryptography is not immediately obvious to an observer. This avoids flagging the traffic as containing high-value data worth storing and attacking with quantum algorithms in the future.

Security Layers: Multiple layers of encryption (IPsec-post-quantum encryption-IPsec) provide defense in depth, making it harder for attackers to decrypt the data even if they manage to break one layer.

My Questions:

Does this approach effectively hide the use of post-quantum cryptography from potential eavesdroppers?

Are there any known vulnerabilities or potential weaknesses in this multi-layer encryption strategy?

Is there a better way to achieve the goal of post-quantum safety while avoiding detection by adversaries?

Any insights or feedback on the feasibility and security of this design would be greatly appreciated.

Answer 1

This protocol is highly inefficient and ineffective and can possibly be redundant against the attacks listed above.

Obfuscation: By wrapping post-quantum encrypted content within regular IPsec, the use of post-quantum cryptography is not immediately obvious to an observer. This avoids flagging the traffic as containing high-value data worth storing and attacking with quantum algorithms in the future.

SNDL (store now decrypt later) pools store all data passed through transport, which should store that too. However, if such a system could selectively store data, it would store data from algorithms that are pre-quantum which are vulnerable to cryptographically relevant quantum computers.

Security Layers: Multiple layers of encryption (IPsec-post-quantum encryption-IPsec) provide defense in depth, making it harder for attackers to decrypt the data even if they manage to break one layer.

If an attacker had the capability to break the first 2 layers of protocol security in trivial amounts of time, the third layer would be practically useless.

Does this approach effectively hide the use of post-quantum cryptography from potential eavesdroppers?

Yes, unless the eavesdropper has a cryptographically relevant quantum computer, however a potential eavesdropper being able to get that data shouldn't affect the security of a secure protocol.

Does this approach effectively hide the use of post-quantum cryptography from potential eavesdroppers?

Yes, however this would not affect the security of the protocol.

Is there a better way to achieve the goal of post-quantum safety while avoiding detection by adversaries?

A better way would be using a HPKE to combine pre-quantum and post-quantum public-key key exchange algorithms and would only require one layer, reducing the about of network transfers. However, this may not even be required, as there is currently a standardization effort by the IPSECME working group to standardize the use of ML-KEM, the post-quantum KEM standardized by NIST, in IPSEC, currently being documented in this Internet Draft.