Mullvad exit IPs are surprisingly identifying
A deep dive into Mullvad, a highly-regarded privacy VPN, reveals that its exit IP assignments are not truly random but deterministically linked to a user's WireGuard key. This subtle technical detail, likely stemming from an unexpected behavior of seeded random number generators, creates a unique cross-server fingerprint. The discovery raises significant privacy concerns about potential user correlation and deanonymization for a service lauded for its anonymity.
The Lowdown
A technical analysis of Mullvad VPN's exit IP assignment mechanism uncovers a surprising design choice: instead of fully randomizing exit IPs per connection, Mullvad deterministically assigns them based on a user's WireGuard key. This behavior, likely an unintended consequence of how seeded random number generators operate with varying bounds, creates a subtle but significant vulnerability.
- The author's script repeatedly changed WireGuard pubkeys and fetched exit IPs for a set of servers, producing data for thousands of pubkeys.
- Despite a theoretically vast number of unique IP combinations possible across servers, only a limited number were observed in practice.
- Further investigation revealed that assigned exit IPs consistently maintained a proportional position (e.g., 81st percentile) within their respective server's IP pool, suggesting the use of a seed-based RNG where a consistent underlying 'float value' is scaled by the pool size.
- This 'float value' effectively acts as a fingerprint, allowing different sessions from the same user (even across various Mullvad servers) to be linked with a high probability, enabling potential correlation of identities.
- For example, a moderator with IP logs could link two accounts to the same individual based on these overlapping float ranges, even if they used different Mullvad servers.
- The author provides a tool to estimate the float range for given IPs and suggests mitigating steps: avoid switching servers more than once per pubkey, and force rotate your pubkey by logging out of the Mullvad app.
This finding exposes a subtle but critical flaw in Mullvad's IP assignment logic that, under certain circumstances, could compromise user anonymity. It serves as a potent reminder that even highly respected privacy tools can harbor unforeseen vulnerabilities from complex technical implementations, underscoring the ongoing challenge of achieving true online anonymity.
The Gossip
Deterministic Debates: Mullvad's Method Mystery
Commenters debated the rationale behind Mullvad's deterministic IP assignment. Some speculated it was a pragmatic choice to prevent a single abusive user from getting all others on a server blocked by major services, thus protecting IP reputation. Others wondered if it was a coding oversight or a misunderstanding of how seed-based random number generators behave.
VPN Efficacy: Snake Oil or Sanctuary?
A common HN debate resurfaced, with one commenter dismissing VPNs as 'snake oil.' This was met with strong counter-arguments highlighting VPNs' benefits, such as transferring trust from an ISP to a high-trust VPN provider, hiding traffic, preventing IP-based tracking, and bypassing geo-restrictions. Some offered a nuanced view, acknowledging that commercial VPNs aren't universally trustworthy but still offer advantages over direct ISP connections.
Privacy Perils: Profiling Potential
The discussion quickly turned to the core privacy implications of the discovery. Commenters noted that deterministic exit IPs enable websites to build persistent user profiles across sessions, effectively shifting the tracking mechanism. The idea that a privacy-focused VPN like Mullvad could inadvertently create such a fingerprinting vector was seen as a significant compromise, potentially facilitating correlation attacks for deanonymization.