Register
News/Arweave: Enabling the Permaweb
Arweave: Enabling the Permaweb

Medium-2021-12-09 21:13:39

The Arweave Protocol is designed to provide decentralized, scalable, and permanent on-chain data storage. Much like how Ethereum is the world’s computer, Arweave can be thought of as the world’s hard drive that never forgets. The protocol uses a combination of innovative blockchain technologies and clever game theory design to create a healthy and sustainable network.

Permanent Storage Through Incentives

Arweave uses a new type of data structure called blockweaves, which iterates on blockchain’s original design. Each block is linked to two prior blocks:

  • The previous block in the chain (similar to traditional blockchains such as Bitcoin); and

  • A block from the prior history of the blockchain (a “recall block”)

Miners must have access to a recall block to mine new ones and gain rewards (Succinct Proofs of Random Access, or SPoRA). Because the choice of the recall block is unpredictable, miners are incentivized to store larger amounts of data, thereby increasing their odds of having access to the correct recall block, allowing them to mine new ones and receive mining rewards. This incentive mechanism also encourages miners to store rarer blocks instead of well-replicated ones, as they compete with a smaller number of miners for the same level of rewards.

Source: Arweave Yellow Paper

Therefore, Arweave takes a probabilistic and incentive-driven approach to ensuring data replicability and permanent storage. The net effect of these incentives is that there is a virtually nil probability that blocks are dropped by the network.

Source: Amber Group. Highlighted in blue approximates the current network state.
Note that the figures above assume that miners store blocks randomly. Because miners are incentivized to store rarer blocks, the actual probability of dropped blocks is lower than reflected.

Pay Once, Store Forever

Arweave users pay a one-time, up-front fee to store their data permanently. This contrasts to several Web2 or Web3 storage platforms, such as Amazon Web Services, Google Cloud, or Filecoin, in which users typically have recurring transactions. We believe Arweave’s model provides a completely different experience — one that is more appropriate for Web3. Whereas customers that subscribe to pay-as-you-go platforms risk being “held hostage” to future price increases or changes of terms, Arweave users can access stored data repeatedly for free.

Not all the transaction fees that users pay go to a miner immediately. Instead, ~86% of them are contributed to an endowment, which is distributed to miners over time to ensure sustainable miner economics (and hence, perpetual storage). The endowment is currently building up its reserves and is not expected to start paying out until Arweave’s permaweb is several times larger than the current surface web.

The sustainability of the endowment model is backed by the assumption of decreasing storage prices over time. Over the past 50 years, the average annual rate of decline in storage costs has been 30.6%. Arweave’s transaction pricing model conservatively estimates that these costs will decline by just 0.5% per annum.

Note that whereas Moore’s Law is decelerating for CPU space, current maximum data density in consumer storage hardware is still far below the density achieved in research or the theoretical maximum limit. Even assuming a rather optimistic 30% annual data density growth rate, it would take over 400 years to hit the theoretical limit. In other words, Arweave’s endowment model is designed to sustain miner economics for centuries.

Leveraging a Proven Game Theory Design

BitTorrent was launched in 2001 and became the largest peer-to-peer file sharing protocol, at one point accounting for over 60% of global internet traffic. Its success is largely due to its elegant game-theoretic design, termed “optimistic tit-for-tat.” BitTorrent nodes optimistically cooperate with other peers but will punish non-cooperative behavior, nudging most actors in the network to act benevolently.

Arweave borrows its design from this proven framework. Each miner in the Arweave network has its proprietary algorithm for ranking other nodes, ranking peers that it perceives to have the highest utility and proportioning its scarce resources to them accordingly. Consequently, Arweave’s network is highly resilient and adaptive to new circumstances.

For example, this March, some nodes in Arweave’s network identified a portion of miners using pooling software, detracting from Arweave’s objective of maximizing data replicability. Subsequently, another set of miners modified their ranking algorithms to deprioritize these miners, thwarting free-loading behavior and improving overall network health.

Bundling: Unlocking Scalability

This September, Arweave released what is effectively a Layer-2 scaling solution: Bundles. With bundling, each Arweave transaction is moved off-chain, grouped with other ones, and then placed back onto the main chain as one large transaction. The upshot is that extremely large data files can be easily uploaded to Arweave’s network. In September, 47GB of data was uploaded in one single block. Bundling also improves transaction certainty and the developer experience. The release of bundling catalyzed step-change growth in total blockweave size, suggesting strong resonance amongst users and developers.

Source: Viewblock

Market Landscape

Arweave typically draws comparison to Filecoin, perhaps the most well-known decentralized storage platform.

Filecoin was founded in 2014 by Protocol Labs, the team behind the creation of the InterPlanetary File System (IPFS). IPFS is the base technology protocol for decentralized data storage that uses content-addressing to identify files. Whereas most files are currently found by where they are (i.e., which server they are stored on), IPFS identifies files by what the data is. With those identifiers, users can share and download files amongst each other. Filecoin acts as an economic layer on top of IPFS to incentivize nodes to host IPFS data.

The economic model between Arweave and Filecoin differs in many aspects. Filecoin operates with a pay-as-you-go model, much like AWS and Google Cloud. And distinct from Arweave, Filecoin primarily offers temporary storage solutions instead of permanent ones.

Moreover, within Filecoin’s protocol, there are thousands of different contracts between users and nodes, each with different terms of storage (such as price, duration, number of replications, etc.) In contrast, there is effectively only one contract offered on Arweave’s platform: data stored forever.

Hence, we do not believe that Arweave directly competes against Filecoin (and similar variants). Their solutions may actually be complementary. In some cases, it will be more cost-effective to store data permanently; in other cases, paying for short-term storage makes more sense.

In fact, it is difficult to find another platform with a similar value proposition as Arweave — IPFS comes closest, but without economic incentives, files can still get dropped from IPFS’ network. For example, Infura’s IPFS pinning service deletes user data if it hasn’t been accessed in six months.

Source: Web3 Index, Viewblock, Siastats, Filecoin, Chronobot, CoinGecko, press releases

Use Cases in the Web3 Stack

Arweave initially saw tepid growth after mainnet launch but gained significant market traction this year with a number of Web3 applications. Below, we outline a number of use cases to show how Arweave found market fit and more clearly illustrate the unique opportunities that the permaweb enables.

Providing Security for NFTs

Purchasers of NFTs are supposedly buying an immutable, permanent object. However, those files are rarely stored on-chain — doing so would be prohibitively expensive for most NFT projects. Instead, smart contracts for NFTs typically store the actual content in a simple Token URI, which points to an internet address where the digital object actually resides.

This poses a problem: most URIs cannot guarantee permanence or immutability. As one example, the World of Women NFT project stores its image files on Amazon Web Services, making it possible to change and delete the files in the future. A survey of the top 22 NFT projects by @pencilflip found that 50% stored their files on centralized servers.

Source: Pencilflip

Even IPFS does not guarantee permanence. CheckMyNFT found several assets from blue-chip artists, including Grimes, Deadmau5, and Steve Aoki, that temporarily went offline despite being hosted on IPFS. As another example, nft.storage purportedly offers data that “will be available in IPFS indefinitely.” However, its terms and conditions caveats that “data will continue to be persisted ad infinitum or .”

Because this issue has become more salient, artists and creators are increasingly using Arweave’s platform for NFT storage, including Beeple and Jay-Z.

“T. The metadata associated with this NFT is higher than current industry standards. The metadata is stored within the smart contract and meets all standards within the Ethereum Improvement Proposals (EIP) adopted by all NFT platforms. 

Sotheby Condition Report on Jay-Z’s first NFT, Heir to the Throne

Outside of images, media platforms such as Mirror (text articles as NFTs) and Pianity (audio NFTs) also host their files on Arweave. As Arweave is uniquely positioned to offer decentralized permanence and immutability, we expect its adoption to continue to grow within the NFT and creator economy.

User Interface — Hidden Point of Centralization

The explosive growth of DeFi illustrated the amount of innovation that can be unlocked by providing users with freely accessible, autonomous, and transparent platforms. However, these platforms face a hidden centralization risk: the front-end user interface. Cloud providers still have the ability to stop and censor dApps, and developers could force front-end changes and “updates” even if most users prefer the older version.

This has already happened. This July, Uniswap removed tokenized stocks from its front-end in fear of U.S. regulations. The smart contracts underneath the user interface still existed, developers just wanted it hidden away. In response, users uploaded Uniswap’s prior UI on Arweave, allowing users to access Uniswap as it existed prior to the censorship. Other platforms, such as SushiswapCompound, and 1inch are also available on the permaweb, minimizing the threat of censorship.

Blockchain Storage

Blockchains are ledgers that perpetually grow. These ledgers are meant to exist forever, but few are incentivized to store old chain data. This issue compounds as the blockchain network grows.

Source: Blockchain.info

The problem is particularly acute for high-performance blockchains such as Solana. Solana has already produced over twice the number of blocks compared to Ethereum, Bitcoin, Polkadot, Algorand, and Cosmos combined.

Instead of designing in-house solutions, developers have turned to Arweave’s technology.

“One of the challenges with building a high-performance smart contract platform is ensuring that the ledger data has a highly fault-tolerant, decentralized storage solution…in a multi-chain future where chains can and should specialize and interoperate. 

- Anatoly Yakovenko, Co-Founder of Solana Labs

Arweave has already partnered with Solana, Cosmos, Polkadot, Avalanche, NEAR, and Skale. The Graph, a blockchain indexing protocol, will also support Arweave, allowing more efficient access to archived blockchain data.

Other Use Cases

We demonstrated only a few examples above, but Arweave’s technology addresses several Web3 and real-world problems, such as financial auditsmedia publishing, and information archiving. In addition, we believe that by creating new features and functionalities, Arweave’s technology will enable new markets and use cases, enhancing the total available market.

Tokenomics

Arweave was released with a genesis supply of 55 million AR tokens. The total maximum token circulation will be 66 million AR tokens. New AR token issuance is similar to Bitcoin’s model, in which coins released are halved at regular intervals, except that Arweave continuously halves its issuance rate whereas Bitcoin has discrete halving events.

Currently, there are a total of ~64 million AR tokens issued. Around 22% of total issued tokens are held by advisors, the project team, or held for future use and development. This suggests that current public holders face a maximum of 24% dilution (22% from the release of current tokens from insiders + remaining lifetime inflation of 2%).

Source: Arweave Yellow Paper, Amber Group estimates

The value of AR tokens is expected to increase with greater demand for storage on Arweave’s protocol. In addition, tokens are continuously taken out of public circulation as the bulk of transaction fees are sent to Arweave’s endowment. As a simple illustration, if Arweave’s total network grows to 10PB and the fee per TB is ~200 AR (current costs are ~480 AR per TB), roughly 1.7 million tokens will be taken out of public circulation. These tokens will eventually be reintroduced into circulation — miners need to sell them down to cover costs of storage — but the pace of re-entrance will likely be gradual and take place over time.

Therefore, Arweave users should expect value accretion to their tokens from both increased demand with fixed supply and retired tokens into endowment from transaction fees.

Closing Thoughts

The team behind Arweave has thoughtfully bootstrapped its network. It introduced Profit Sharing Tokens that allow developers to earn a stream of “micro-dividends” from the apps they build on Arweave, incentivizing founders to build applications on top of Arweave. It also regularly hosts the Open Web Foundry, a six-week program to help entrepreneurs bring their ideas to market on Arweave’s platform. All of this contributes to a burgeoning Arweave ecosystem.

Source: Verto Protocol

It is still early days; execution and adoption risks abound. Nevertheless, Arweave’s mission is ambitious. If successful, it could act as the de facto decentralized data platform for Web3 applications. More broadly, it could also bring us one step closer to building a truly permanent library of human knowledge.

Further Resources

Disclaimer

The information contained in this post (the “Information”) has been prepared solely for informational purposes, is in summary form, and does not purport to be complete. The Information is not, and is not intended to be, an offer to sell, or a solicitation of an offer to purchase, any securities. The Information does not provide and should not be treated as giving investment advice. The Information does not take into account specific investment objectives, financial situation or the particular needs of any prospective investor. No representation or warranty is made, expressed or implied, with respect to the fairness, correctness, accuracy, reasonableness or completeness of the Information. We do not undertake to update the Information. It should not be regarded by prospective investors as a substitute for the exercise of their own judgment or research. Prospective investors should consult with their own legal, regulatory, tax, business, investment, financial and accounting advisers to the extent that they deem it necessary, and make any investment decisions based upon their own judgment and advice from such advisers as they deem necessary and not upon any view expressed herein.


View Original