Skip Navigation

Clover vs Dandelion++ | Network-Layer Privacy in Monero

Clover vs Dandelion++ | Network-Layer Privacy in Monero NOTE: SEE THIS IN MONEROKON IN JUNE 21 (06/21/2025)!

I've been reading about Clover and Dandelion++, two techniques aimed at improving privacy on the network layer specifically, hiding the origin IP of a transaction in systems like Monero.

Here’s a quick comparison and breakdown:

🌼 Dandelion++

How it works:

Uses a two-phase propagation system:

Stem phase: The transaction is passed through a few nodes in a linear, random path.

Fluff phase: It’s then broadcast normally (flooded).

Privacy benefits:

Helps hide the source IP by making it harder for attackers to pinpoint the origin based on initial propagation.

Works well against local or partial adversaries (those controlling a few nodes or monitoring part of the network).

Already implemented in Monero.

Limitations:

Vulnerable to global adversaries (e.g., an attacker that can monitor large parts of the internet).

Susceptible to timing analysis and some advanced correlation attacks.

🍀 Clover

How it works:

Proposes a more advanced, probabilistic routing system.

Introduces random delays, mixing, and adaptive path selection using buffers and stochastic rules.

Makes the traffic pattern statistically indistinguishable from honest noise.

Privacy benefits:

Much stronger protection against global passive adversaries.

Defends against timing attacks and traffic analysis far better than Dandelion++.

Higher entropy in transaction routing paths.

Limitations:

More complex to implement.

Likely introduces higher latency due to delays and buffering.

Not yet implemented in Monero—still a research prototype (see more in MoneroKon).

Conclusion: Dandelion++ is a practical and effective step forward for Monero’s network privacy, but Clover shows promise for the future, especially if we want to defend against more powerful, surveillance-level attackers. It’s a trade-off between deployment complexity and stronger anonymity guarantees.

Reddit

Comments

1