Vegas: Reclaiming Digital Sovereignty through Restricted P2P Social Networking

TL;DR

Vegas is a radical P2P social network design that rejects the "global discovery" features of Facebook in favor of extreme privacy. By limiting a user's view strictly to their immediate friends (ego network) and using unique encryption keys for every single relationship, Vegas ensures that neither service providers nor malicious peers can map the broader social graph.

Context: The Failure of Centralization and "Open" Decentralization

Current Online Social Networks (OSNs) treat user data as a commodity. Even decentralized attempts like Safebook or Diaspora often fall short:

• Safebook relies on DHTs (Distributed Hash Tables) which are vulnerable to traffic flow analysis.
• Diaspora uses "pods" that still allow for unsolicited friendship requests, leading to "Social Network Pollution."

The authors of Vegas argue that Informational Self-Determination—the ability to fully control the storage and dissemination of one's data—is impossible if the system allows strangers to browse the social graph.

Methodology: High-Security Architecture

Vegas introduces several novel concepts to solve the privacy-functionality trade-off.

1. Link-Specific Cryptography

Unlike standard PKI where you have one public key for the world, in Vegas, if you have 50 friends, you have 50 different public keys.

• Benefit: No two friends can "collude" to prove they are talking to the same person based on a public key ID.
• Anonymity: The fingerprint of the key acts as an anonymous identifier that only the specific recipient can recognize.

2. The Exchanger Model

Vegas doesn't care how a message gets there. It uses "Exchangers"—abstract asynchronous queues.

• Multi-channel: A message might be sent via Email, XMPP, or even split into SMS fragments.
• Traceability Protection: By switching channels (e.g., sending sensitive data via a Twitter "pinwall" while sending the decryption key via SMS), users can evade traffic analysis by ISPs.

Fig 1: The Vegas Architecture, showing the separation between Clients, Exchangers (for transport), and Datastores (for persistence).

3. Friendship via Physical Trust

Vegas prevents "Social Network Pollution" by requiring Out-Of-Band (OOB) friendship establishment. You cannot "search" for a user. You must meet them in person and scan a QR Code, or be introduced via a "Coupling" protocol by a mutual trusted friend.

Fig 2: In Vegas, the user's view is strictly limited to their immediate neighbors. The rest of the network is invisible.

Intelligent Channel Selection

On mobile devices, choosing a communication channel is a trade-off between cost, battery, and privacy. Vegas implements a Weighted Arithmetic Mean model:

$W(k)=\frac{{\sum }_{c\in C}(w(c)\ast f_c(k))}{{\sum }_{c\in C}w(c)}$

This allows the app to automatically switch from XMPP (fast, but high battery/traceability) to Email or SMS depending on whether the user is on WLAN or 3G, or if they require maximum anonymity.

Table 1: Qualitative comparison of different exchanger channels used in the Vegas prototype.

Critical Insight: The Cost of Privacy

Vegas intentionally sacrifices "virality." You cannot find "friends of friends" unless they are explicitly introduced. While this limits the growth of the network, it solves the fundamental security flaw of modern OSNs: the visibility of the social graph. In Vegas, the social graph is not a single entity; it is a collection of private, overlapping ego networks.

Conclusion & Future Work

Vegas proves that a high-security OSN is possible if we are willing to return to a trust model based on real-world interactions. Future work targets "Social Search" relaxations—allowing users to query the network beyond their ego network using encrypted, anonymous forwarding, ensuring that functionality can be added back without compromising the core privacy foundation.