[IEEE Network] WhozThat: Breaking the Ice with Mobile Social Ecosystems

TL;DR

WhozThat is a fundamental framework for Mobile Social Networks (MoSoNets) that answers the universal question: "Who is that person near me?" By combining local wireless ID broadcasting with cloud-based social profile retrieval, it transforms passive physical spaces into "socially-aware" environments. Its most striking application? A jukebox that knows exactly what song a room full of strangers will collectively enjoy.

Problem: The Physical-Digital Social Gap

Despite the explosion of online social networks like Facebook and LinkedIn in the late 2000s, a massive gap remained: Local Context. You might have 500 "friends" online, but you are a stranger to the person standing two feet away in a bar.

Prior works like CenceMe focused on uploading local sensor data to the cloud for remote friends to see. WhozThat flips the script. It aims to import the digital social graph into local physical proximity to lower the barriers to human discourse. The researchers identified that the main technical hurdle wasn't the data itself, but the binding of identity to a dynamic, mobile location.

Methodology: The Two-Phase Protocol

The core of WhozThat is a surprisingly elegant two-step protocol designed for efficiency and platform agnosticism:

1. The Meet-and-Greet Phase: Users' smartphones periodically broadcast a "Social Handle" (a pointer/ID) via local wireless links like Bluetooth or WiFi. This is lightweight and avoids the high energy cost of constant GPS usage.
2. The Context Creation Phase: Once a phone picks up a nearby handle, it uses its wide-area connection (3G/WiFi) to query social network APIs and fetch relevant public profile data (hobbies, music, mutual friends).

Figure 1: The dual-phase workflow where Anne and Dave’s devices exchange handles locally before fetching cloud data.

The "Ecosystem" Vision

The researchers didn't just want a social app; they wanted an ecosystem. They proposed three expansion tiers:

• Context-Aware Local Services: Environments (like bars or stores) that listen for handles and change their behavior.
• Multihop Mesh Networking: Allowing IDs to travel through a crowd, even if the source is out of direct radio range.
• Gateway Services: Offloading the heavy lifting of data mining and profile processing to edge servers.

Implementing the Context-Aware Jukebox

To prove this wasn't just theory, the team built WZPlaylistGen. This application acts as a "Social Jukebox."

Figure 2: Prototype implementation showing the social profile retrieval on actual 2008-era mobile hardware.

The system listens for WhozThat handles over Bluetooth, hits the Facebook API to find those users' "Favorite Artists," and then uses the Audioscrobbler API to generate a collaborative playlist. If three people in the bar love Bob Dylan, the bar starts playing Dylan. It is the ultimate "Ice Breaker," providing a natural conversation starter based on shared interests.

Critical Analysis: The Privacy Elephant in the Room

While the technical implementation is sound, the paper honestly tackles the "Creepiness Factor."

• Security: How do you prevent "spoofing" handles? The authors suggest encrypted MACs with nonces, but the broadcast nature of the system remains a target for "snooping."
• Privacy: Currently, the system relies on existing social network permissions. However, the authors admit that a "friendship-only" disclosure model is too restrictive for a party where the whole point is to meet non-friends. They propose a future "Automated Privacy Policy" that adapts disclosure based on the social context (e.g., more info shared at a mixer, less at a subway station).

Conclusion & Future Outlook

WhozThat effectively pioneered the concept of the "Electronic Caption"—a discreet introduction provided by our devices before we even speak. While the specific platforms (MySpace, JME) have evolved, the core logic of local ID broadcast + cloud context retrieval remains the blueprint for modern proximity-based services and wearable tech.

Takeaway: The next evolution of the social web isn't on our screens; it's in the air around us, mediated by devices that understand the collective identity of a room.