Filecoin’s cryptoeconomic model

Public blockchain networks rely on a combination of cryptography and incentive structures that allow the system to work without centralized control. Cryptography makes some harmful actions virtually impossible, while game theory prevents others.

The discipline that focuses on this set of mechanisms is called cryptoeconomics, and a well-structured cryptoeconomic system can determine whether a network is valuable, self-sustaining, or struggling to get out of trouble.

Today, we are excited to provide an overview of the cryptoeconomic structures that will enable Filecoin to create that foundation. Filecoin uses a combination of existing cryptoeconomic structures and many new ones to meet the unique requirements of the network.

If you’re interested in the long-term success of Filecoin as a storage market, you should find these mechanics encouraging. While we’re still doing a lot of testing to finalize the parameters, we’re excited to share these details about how the system works.

There are five main players in the Filecoin economy:

The role of clients in the Filecoin economy is simple but critical: they use FIL to pay for the storage and retrieval of files. Filecoin's cryptoeconomic mechanism is ultimately designed to meet the file storage and retrieval needs of clients.

Developers are the bridge between clients and the Filecoin network, building applications to improve client access to Filecoin’s storage and retrieval markets.

The retrieval service acts as a content delivery network or caching layer to ensure that files stored on Filecoin are efficiently stored, and clients can pay for retrieval in tiny increments using payment channels.

Storage miners obtain Filecoin services in three main ways:

Most of Filecoin’s cryptoeconomic structure is designed to ensure that storage miners, as well as the Filecoin network itself, can reliably and sustainably meet the needs of customers.

Token holders provide a bridge between the Filecoin economy and other markets, they provide liquidity to miners and clients, and make it easier to build real businesses on Filecoin.

Filecoin storage miners receive block rewards based on their storage power, which is a measure of how much effective storage a Filecoin storage miner provides to the network. At a high level, miners earn storage power by accepting files from clients, publicly committing to store them for a period of time, and repeatedly proving that they are actually storing them over time.

Economically, you can think of storage power like hash rate in a proof-of-work blockchain like Bitcoin. A miner receives a reward proportional to her storage power, not the total storage power provided by all miners.

For example: a miner maintaining 1 PB of storage capacity when the total network size is 100 PB will win, on average, 1% of the block reward over time.

Filecoin needs to become a reliable and sustainable protocol in the long term. In a traditional proof-of-work blockchain, it doesn’t matter if a given miner goes offline for a short time (or forever) because any miner can add new blocks to the chain.

Filecoin miners also provide irreplaceable useful services to their clients, so clients need strong guarantees that each storage miner will continue to operate. Therefore, Filecoin includes many mechanisms to make high-quality services an economically reasonable choice for storage miners and to encourage storage miners to promote the long-term health of the network.

Many blockchains reward tokens based on a simple exponential decay model, where the block reward is highest at the beginning and miner participation is lowest, so mining generates many tokens per unit of work early in the network's lifecycle and then quickly disappears.

In many cryptoeconomic simulations, it is clear that simple exponential decay models encourage unhealthy short-term behavior around network launches.

Specifically, this will incentivize storage miners to complete the closed phase of mining as quickly as possible by overinvesting in hardware. Since Filecoin's rewards per unit of storage power will be much larger in the early days of the storage network, we predict that it will be profitable to exit the network after exhausting these early rewards, even if this results in the loss of customer data.

This would harm the network: clients would lose data and be unable to access long-term storage, while miners would have little incentive to improve the network.

To encourage consistent storage enablement and long-term storage investment, rather than just fast sealing, Filecoin introduces the concept of a network baseline. Instead of minting tokens based purely on elapsed time, block rewards scale as total storage capacity on the network increases. This preserves the shape of the original exponential decay model, but softens it in the early days of the network. Once the network reaches the baseline, the cumulative block rewards posted are the same as in a simple exponential decay model, but at smaller network sizes, part of the block reward is delayed. The overall result is that Filecoin’s rewards to miners are more aligned with the utility they and the network as a whole provide to clients.

Storage miners add fixed amounts of new data to the network called sectors. When a storage miner adds a sector, he commits to storing it for a certain period of time, which gives clients more certainty that the stored data will be available when they need it.

In some cases, miners may agree to a storage agreement and then wish to abandon it later due to increased costs or other market dynamics. Clients need reliable storage, and a system where storage miners can freely or cheaply abandon files will drive clients away from the Filecoin network due to severe data loss and poor service quality.

Filecoin miners work with clients in a two-sided data market, and to ensure that all incentives in the partnership are properly aligned, Filecoin penalizes miners who fail to store files within the promised period.

The penalty is imposed on the pool of Filecoin collateral provided by each storage miner (called locked funds), which includes a small amount of upfront FIL, as well as a large amount of tokens taken from the block rewards received by the miner.

If a miner goes offline for a day, she will lose some money due to locked funds; if a miner stops proving the storage of documents completely, he may lose all documents. If the miner successfully completes the full term of the storage contract, he can recover all locked funds after a short lock-up period.

Using block rewards to collateralize miners is another improvement gained from our cryptoeconomic simulations. Early models required miners to post large upfront collateral for all their storage. However, this would create a Filecoin shortage at launch, when the total supply of tokens would be reduced and many miners would attempt to join the network. By shifting some of the upfront costs to later block rewards, we are able to lower the cost of entry for miners.

Filecoin clients and storage miners operate in a dynamic system where it is impossible for storage demand to exactly match supply at any given time.

Early Filecoin builds forced storage miners to choose between two false options: waiting for client deals to be fulfilled (which meant the miners’ hardware would sit idle), or storing useless data to earn block rewards (which reduced the network’s ability to store useful data).

Filecoin introduces several mechanisms to ensure that miners who store client data can also maximize the FIL they earn.

When a miner finds himself without client data to store, he is ready to provide storage for clients thanks to Filecoin's committed capacity sectors (proven but empty sectors) that show he does not need to keep his hardware idle. Committed capacity sectors verify that the miner has space available for client data, but can be upgraded to store client data at any time. This mechanism alleviates the need for miners to choose between block rewards and client data - he can store committed capacity sectors immediately to start earning block rewards, and then upgrade them to store client data at any time in exchange for transaction fees.

Committed capacity sectors can increase miners' incentives to store client data, but they don't solve the problem completely. Storing real client files adds some operational costs to storage miners. In some cases, for example, if miners' rewards for the value of blocks far exceed transaction fees, miners may still choose to ignore client data completely and just store data they generate themselves to increase their storage capacity rewards as quickly as possible. This will keep Filecoin usage low and limit clients' ability to store data on the network.

Filecoin solves this problem by introducing the concept of verified data. Sectors with verified data will receive more storage capacity and thus more block rewards than sectors without data. This provides storage miners with additional incentives to store customer data.

Verified data is data stored by validating clients, which in turn are verified by a decentralized network of validators. Verification does not imply scarcity - it is very easy for anyone with real data to obtain it and store it on Filecoin. Even though validators can arbitrarily distribute verified data to simplify the registration process, the overall effect should be to significantly increase the proportion of useful data stored on Filecoin.

A decentralized storage network that lets anyone in the world store or retrieve files is an amazing tool. However, the Filecoin network is only useful to the world if there are applications built on it. The good news is that there are already many amazing applications being built on Filecoin, and more will come between now and the mainnet launch.

Filecoin was built from day one to reliably and efficiently store the world's most important data. To improve reliability, the protocol provides customers with unlimited flexibility to store redundant copies of files using different miners and to verify that the only copy is actually stored.

Unlike centralized cloud storage services, which back up data in a way that clients cannot change or verify, Filecoin allows clients to easily express their preferences for reliability and cost.

Filecoin's off-chain data transfer mechanism is another example of the protocol's support for a variety of client needs. Many datasets are so large that transferring them over the internet is slow and expensive.

As Filecoin mainnet launch approaches, you can expect to hear more about applications that are already using the network’s unique capabilities to store large amounts of critical data.

Filecoin’s cryptoeconomics are designed to ensure that the network can meet the long-term needs of its customers, and after thousands of hours of discussion, modeling, and experimentation we believe these mechanisms will lay a solid foundation for the launch of the Filecoin mainnet.

The end result of these efforts, along with the amazing work of Filecoin’s research and engineering teams, will be a public blockchain powered by useful work, a reliable cloud storage service, and a thriving decentralized economy.

In the coming weeks, we will conduct final simulations and modeling to determine the final cryptoeconomic parameters. We look forward to independent analysis from other groups and to the community's contributions to making the Filecoin network more useful and valuable to the world.