Technology

Frog2Frog is a distributed overlay network built on top of existing CDN infrastructure.

At its core:

• The CDN remains the origin, fallback, and authorization anchor
• Each viewer device runs a lightweight Frog2Frog Agent
• Devices consuming the same stream form a dynamic swarm

A swarm is a content-scoped group of devices that exchange video segments directly with one another. It functions as a temporary, distributed micro-CDN dedicated to a specific live stream.

The system is therefore hybrid:

• Centralized control and authorization
• Decentralized execution and distribution

Network formation is deterministic and content-driven.

When a stream starts:

1. The node computes a unique identifier of CDN URL, tokens, or SSAI mutations.
2. Node discovery proceeds via gossip-based neighbor exchange.
3. Devices consuming the same identifier converge into the same swarm.

No central registry is required after bootstrap.

Swarm membership evolves organically and self-heals under churn.

As audience grows:

• Each new participant adds both demand and upload capacity.
• Aggregate delivery capacity increases proportionally.

Scalability becomes an emergent property of the network.

Security is embedded directly into the overlay.

Each node derives its identity from a blockchain-grade pairing-based key pair.

PeerId is cryptographically bound to its public key, enabling:

• Compact signature aggregation
• Efficient multi-party verification
• Strong Sybil resistance

To preserve CDN authorization guarantees, Frog2Frog relies on server-side watermarking and session-aware token management to control and trace content access.

Frog2Frog continuously reorganizes its overlay.

Peer selection is AI-driven and based on real-time metrics:

• RTT
• Throughput
• Bandwidth estimation
• Reliability scoring
• Churn indicators ...

A learning-based model computes optimal neighbor sets to maximize aggregate throughput and minimize congestion.

Topology refinement occurs through distributed swap consensus:

• Nodes propose local rewiring.
• Gains are evaluated.
• Atomic transitions are committed without central coordination.

Segment delivery is opportunistic:

• For each segment, nodes are ranked dynamically.
• The best source,  node or CDN, is selected in real time.
• Instant fallback prevents playback degradation.

The network learns and stabilizes itself continuously.

Although execution is decentralized, operations remain centralized and deterministic.

A multi-channel telemetry architecture streams:

• Debug metrics
• Billing attribution
• Delivery verification
• Network health indicators

Operators access real-time dashboards showing:

• Swarm topology evolution
• Offload ratios
• Reliability anomalies

Remote configuration allows dynamic tuning of:

• Peer selection thresholds
• Verification sampling rates
• Transport preferences
• Feature activation

No redeployment is required for parameter changes.

Frog2Frog is powered by a unified core engine deployed across:

• Linux
• Android
• iOS / macOS
• WebAssembly (browser runtime)

Two integration models are supported:

HTTP Proxy Mode

• Transparent interception
• No player modification required

Custom Segment Loader API

• Programmatic control
• Advanced scheduling and retry strategies

Both modes preserve full compatibility with cryptographic verification and watermark workflows.