The AT Protocol: Introduction & Understanding

The AT Protocol: Introduction & Understanding

What is the Authenticated Transfer Protocol (AT Protocol)?

The Authenticated Transfer Protocol (ATP) is a protocol for large-scale distributed social applications. It allows users to securely exchange data in signed data repositories and migrate their account to a new Personal Data Server (PDS) without the server's involvement. ATP was designed to achieve scale, high degree of user-choice, and freedom to produce user interfaces independently of the servers.

How does it work?

Identity

Users in the ATP are identified by domain names, similar to email addresses. These domain names map to cryptographic URLs that secure the user's account and their data. This allows users to maintain control over their data and protect their privacy.

Data Repositories

User data is exchanged in signed data repositories. These repositories are collections of records that include posts, comments, likes, follows, media blobs, and other types of data. The repositories are secured with digital signatures, making it virtually impossible for anyone to tamper with the data. This ensures that users have complete control over their data and can trust that it has not been modified without their consent.

Federation

The ATP syncs the repositories in a federated networking model. Federation is used to ensure that the network is convenient to use and reliably available. Commands are sent between servers using HTTPS + XRPC. This means that servers can communicate with each other over the internet in a secure and efficient manner.

Interoperation

A global schemas network called Lexicon is used to unify the names and behaviors of the calls across the servers. Servers implement "lexicons" to support featuresets, including the core ATP Lexicon for syncing user repositories and the Bsky Lexicon to provide basic social behaviors. This makes it easy for software from different organizations to understand each other's data, removing the need to exchange rendering code while browsing content.

Achieving Scale

The ATP distinguishes between "small-world" and "big-world" networking. Small-world networking encompasses interpersonal activity, such as mentions, replies, and direct messages, as well as syncing datasets according to follow graphs. Big-world networking, on the other hand, includes large-scale metrics, such as likes, reposts, and followers, content discovery algorithms, and user search.

Personal Data Servers (PDS) are responsible for small-world networking, while indexing services crawl the network to provide big-world networking. The small-world/big-world distinction is intended to achieve scale as well as a high degree of user choice.

Algorithmic Choice

As with web search engines, users are free to select their indexers. Each feed, discovery section, or search interface is integrated into the PDS while being served from a third-party service. This gives users the freedom to choose the algorithms that work best for them, rather than being locked into a single algorithm chosen by the service provider.

Account Portability

The ATP's goal is to ensure that a user can migrate their account to a new Personal Data Server without the server's involvement. User data is stored in signed data repositories and verified by Decentralized Identifiers (DIDs). DIDs are essentially registries of user certificates, similar in some ways to the TLS certificate system. They are expected to be secure, reliable, and independent of the user's PDS.

Each DID Document publishes two public keys: a signing key and a recovery key. The signing key is entrusted to the PDS so that it can manage the user's data, but the recovery key is saved by the user, such as a paper key. This makes it possible for the user to update their account to a new PDS without the original host's help.

Speech, Reach, and Moderation

ATP's model is that speech and reach should be two separate layers, built to work with each other. The “speech” layer should remain neutral, distributing authority and designed to ensure everyone has a voice. The “reach” layer lives on top, built for flexibility and designed to scale.

The base layer of ATP (Personal Data Repositories and Federated Networking) creates a common space for speech where everyone is free to participate, similar to the web where anyone can put up a website. The Indexing services then enable reach by aggregating content from the network, similar to a search engine.

The End Notes

In conclusion, the Authenticated Transfer Protocol (ATP) is a decentralized protocol designed for large-scale distributed social applications. It provides a federated networking model that ensures convenience and reliability, allowing for the exchange of signed data repositories between servers. With the use of a global schemas network called Lexicon, the protocol unifies the names and behaviors of calls across servers, enabling software from different organizations to understand each other's data.

ATP distinguishes between small-world and big-world networking to achieve scalability and provide a high degree of user choice. It also allows for algorithmic choice, where users can select their indexers, and provides account portability, allowing users to migrate their account to a new Personal Data Server without the server's involvement.

The speech and reach layers of ATP are designed to work with each other, creating a common space for speech where everyone is free to participate, and enabling reach by aggregating content from the network. With the use of DIDs and backup syncing, ATP ensures secure and reliable user data.

As decentralized systems become increasingly important in today's digital landscape, protocols like ATP provide a reliable and secure way for users to participate in social applications while maintaining control over their data.