The Composable Web3 Data Network

A few months ago at the 2021 Multicoin Summit, I gave a keynote presentation highlighting composability as the most important development in crypto in 2022. Although that presentation specifically focused on composability of financial assets on the Solana blockchain, financial composability is not the only form of composability. There’s an even larger opportunity for composability: data composability.

The foundational substrates of the crypto ecosystem are Layer 1s. These networks are broadly called Layer 1s because they each have a set of sovereign nodes that come to consensus about state at some cadence. However, they are functionally distinct based on what type of state they track, how they achieve their consensus, and the types of applications they can support. There are a few major categories of L1s:

1. Asset Ledgers — Bitcoin, Ethereum, Solana, Polygon, Flow, Near, Avalanche, Terra, etc.
2. Storage Ledgers — Filecoin, Arweave, Sia, etc.
3. Data Ledgers — Ceramic
4. Application-Specific Ledgers — Osmosis, Helium, Thorchain, Sommelier, Auth Network, etc.

We have firmly believed for years (see here, here, and here) that the Solana blockchain will become the dominant crypto-native asset ledger. We believe that Ceramic will become the dominant data ledger, and today we’d like to announce that we co-led a $30M investment in Ceramic alongside Union Square Ventures. In addition to USV, the round included Coinbase Ventures, CoinFund, Collabfund, Dapper Labs, DCG, Edge and Node, Figment, Hashed, Jump Crypto, Metacartel Ventures, Not Boring Capital, Northzone, P2P Capital, Placeholder, Protocol Labs, Reciprocal Ventures, Variant Fund, Venrock, and more than 50 crypto angels and community members.

Scarcity and Abundance

Asset ledgers like Ethereum and Solana do one very simple thing: keep track of who has how many tokens (both fungible and non-fungible tokens). That’s it. The net output of these systems is deceptively simple.

Framed another way, asset ledgers manage scarcity. For any given asset at a single point in time, there must be a fixed number of tokens. In every transaction—payments, trades, borrows, etc.—someone decrements their token balance, and another person increments their token balance.

However, the general arc of the Internet over the last 30 years has not been about scarcity. In fact, it has been the opposite. It has been about abundance. More of everything—more data, more pictures, more computers, more servers, more friends, more followers, more connections, more consumption, more sharing, and so forth.

In scarcity-centric applications, as some users gain coins others must lose coins. But in abundance-centric applications, users can create an infinite amount of data and content. You can write an unlimited number of Tweets and follow an unlimited number of accounts. This key distinction necessitates the need for a new type of Layer 1 that can handle these types of data-centric use cases and transaction volume.

What Solana is to scarcity-centric web3 applications (who has how many tokens), Ceramic will be to abundance-centric web3 applications (who authored which tweets). If there is ever going to be a decentralized version of Twitter, dTweets will be hosted on Ceramic.

With the explosion of NFTs, web3 social apps, crypto games, and more, there is now ample developer interest to bootstrap a composable, decentralized database. Today there are thousands of developers who’ve already built over 400 applications on Ceramic, making Ceramic far and away the leading decentralized data network.

System Overview

All ledgers—asset ledgers, storage ledgers, data ledgers, etc.—must achieve three properties concurrently:

1. Scale — Without scale, software is just a toy.
2. Composability — As more data, state, and functions are added to a decentralized ledger, they increase the breadth and depth of the substrate on top of which new applications can be built. Composability is the ultimate network effect.
3. Logical centralization — As systems logically fragment (usually to support scale), they become increasingly complex and difficult to build on. While the core benefit of composability is compounding network effects, breaking logical centralization to enable scale prevents developers from actually realizing this benefit. We have emphasized the importance of logical centralization for years.

SCALE

One reason that Solana has been successful is that it achieves tens of thousands of transactions per second (and will hopefully scale to millions in the coming years) for financial applications. If Ceramic is going to concurrently host decentralized versions of Twitter, Facebook, Reddit, Snapchat, TikTok, etc., Ceramic must scale to hundreds of millions of transactions per second. This is an entirely different level of scale.

To achieve that, Ceramic makes a few very opinionated decisions on the data structure of its ledger. The most important is that in Ceramic, there is no notion of state that can be shared between people (e.g., a token that can be transferred, or an AMM pool). Every piece of state is owned only by the person who created it, and no person can modify anyone else’s state (though any user can link to a piece of state owned by another user).

A useful mental model of Ceramic is that every user has a collection of JSON documents, called Streams, that only they as the owner of those documents can modify. The content stored in each document is arbitrary, and can reference content in anyone else’s documents. Note that this does not preclude compute. Developers can write functions, called streamcode, that define how these documents can be updated and what actions they perform upon each new update. For example, for a Twitter-like application, developers can define a userTotalTweetCount function and increment that as users publish each Tweet to their Ceramic stream.

The beauty of this opinionated architecture—that untangles state between users—is that the system scales horizontally very cleanly. You can imagine a world in which user accounts 1 - 1,000,000 are replicated on one set of Ceramic nodes, and user accounts 1,000,001 - 2,000,000 are replicated on another, etc. Theoretically, the network can shard all the way down to each individual user if needed without breaking composability. In order to ensure state verifiability and composability between user shards, Ceramic relies on a merkle tree data structure that aggregates transactions across all users, allowing any user to verify the integrity of anyone else’s JSON documents at any time.

COMPOSABILITY

Ceramic achieves cross-application data composability primarily through its use of a novel abstraction, called data models, which unifies how similar types of applications store and retrieve state from each individual user on the network. For example, you can imagine that every decentralized Twitter implementation would run on a few shared data models: one for each user’s tweets, one for their social graph, one for their DMs, etc. By adopting the same underlying data models, applications are able to natively interoperate on the same data.

In a way, you can compare Ceramic’s use of data model standards to the use of token standards for asset ledgers. On Ethereum, for example, the introduction of the ERC-20 fungible token and ERC-721 non-fungible token standards have given rise to entire ecosystems of tokens and financial applications that natively interoperate. Ceramic brings this same concept to data.

Ceramic takes a community-driven approach to creating these data models, allowing any developer to easily define, share, and reuse their models with other developers in the ecosystem. As more data models are created by the community, we will see a continuous expansion in the quantity and variety of applications that are built with composable data.

Composability done this way also makes the developer experience better. Building an application on Ceramic looks like browsing a marketplace of data models, plugging them into your app, and automatically gaining access to all data on the network that’s stored in these models. Using Ceramic, developers won’t need to worry about bootstrapping their application with their own siloed users and data. The rate of compounding innovation across developers is going to accelerate dramatically.

L.F.G.

The Ceramic network is live and the SDK is available here. We have already invested in a handful of companies who are building on top of Ceramic.

The Ceramic network has been in beta since June 2021, and is being hosted by a handful of early supporters. Over the coming months, the network will decentralize to an increasing number of hosts and become fully permissionless so that anyone can host content in the Ceramic network. In order to transition the network to permissionless hosting, Ceramic will launch a token, and that token will be used as the economic layer to incentivize people to host the Ceramic network all over the world. This is somewhat analogous to IPFS and Filecoin.

One of my favorite things about Ceramic is that it unlocks new features for developers who are building on asset ledgers today. Developers who are building on any of the major asset ledgers—Ethereum, Solana, Polygon, Avalanche, Near, Luna, Flow, etc.—can simultaneously leverage Ceramic for data-centric functions to make their applications better. Through Ceramic’s flexible DID-based account system, Ceramic naturally interfaces with the private keys that users of any of the major asset ledgers are already using. Because of this, I expect to see developers who are building on all the major asset ledgers to also adopt Ceramic for parts of their applications.

As thousands of developers launch applications on Ceramic this year, the amount of composable state in the Ceramic network is going to explode, providing a vibrant data substrate to create continually higher-order applications.

If that excites you, Ceramic is hiring.