Arabic / Chinese / Japanese / Latvian / Russian
As modern blockchains grow in capacity by orders or magnitude, so does the data they produce. Bitcoin has been around for over 12 years and has a total size of 370 GB; Ethereum, 6 years old, 1039 GB, up 95% in the last year. Solana, a year and a half old, with a throughput that can kick out 4 petabytes of data annually.
The more capacity a chain has, the faster it comes up against the problem of scale. Not the traditional problem of scale around keeping gas fees cheap and transaction speed high, but the problem of being able to actually store and recall all of that data somewhere decentralized, verifiable, and accessible for all.
Trying to strike the balance between scalable, secure, and decentralized is a problem as old as cryptography. As we’ll look at in this article, different blockchains address it in different ways, but the solution may be as simple as using Arweave as a ‘layer 0’, validated by a layer of nodes on top to ensure data correctness.
Before getting into how Arweave fits in, we’ll take a deeper dive into the trilemma of decentralization and the hurdles blockchains face when storing their own historical data.
Defining layers 1 and 2
A layer 1 is another term for a blockchain’s main architecture – the ‘base layer’. Ethereum, Bitcoin and Solana are all examples of layer 1s. Of course, the term wouldn’t need to exist if there weren’t more layers to consider.
Layer 2s are separately integrated networks, usually built to solve problems of scale. Optimism, for example, is an Ethereum layer 2 that solves the layer 1’s notoriously high gas fees by sending transactions as batches, paying gas once, but splitting the fee across the many participants. It also solves for scalability, since the layer 1 is only able to process 13 transactions per second despite seeing demand for an enormous amount more.
Layer 2s also exist for Arweave. The Bundlr network is an alternative way to upload batches of data to the permaweb which comes with an added guarantee that no transaction will be dropped. Scaling issues with layer 1s can usually be solved by elegant layer 2 designs. But storage issues are another matter.
The problem of storage
Blockchain size only ever goes up, and an architecture that makes it necessary for miners and validaters to store the full ledger is going to find itself becoming centralized as node operators without petabytes of storage space are priced out of the game.
See for example this post on the topic of ‘CeDeFi’, describing why it is necessary that Binance Smart Chain has a certain degree of centralization:
“Binance Smart Chain increases roughly at a 10-times faster rate than the state on the Ethereum blockchain. Currently, with an average block size of 40,000 bytes, Binance Smart Chain grows by around 1.15 GB per day which is around 420 GB per year. After a couple of years, this of course eliminates most of the consumer-grade hardware.”
The Binance solution was to commission and control its own network of validators. A Messari report from April 2021 estimates that the vast majority of the network is governed by organizations which are either Binance partners, or “presumably Binance-managed or influenced”.
Some might overlook the influence Binance Chain has over Binance Smart Chain's validator set.
BSC has 21 active validators, making it more centralized than most platforms.
This validator set is determined daily by Binance Chain, a network managed by just 11 validators. pic.twitter.com/UeWDj0o1DY
— Wilson Withiam (@WilsonWithiam) April 12, 2021
Binance sacrificed decentralization for speed and affordability in a move that the Ethereum Foundation and die-hard decentralizooors would vehemently resist. It earned Binance Smart Chaim (BSC) plenty of harsh criticism.
Offloading the job of storage onto a finely-tuned network of gargantuan centralized servers is one way to solve the storage problem, but in Binance’s case – with only a handful of nodes that can provide transaction history, who validates the validators?
Arweave is Solana’s layer 0
Out of any layer 1 we’ve discussed here, the question of storage is perhaps most pertinent for Solana. If 420 GB per year sounded hard to manage, try 4 petabytes. That’s 4,000,000 GB of transaction history generated every year if Solana is running at maximum capacity.
“[Since its inception in March 2020] the Solana blockchain has produced more than two times the number of blocks as Ethereum, Bitcoin, Polkadot, Algorand, and Cosmos blockchains combined”- Solana Labs Medium, Dec 8 2020.
Since it would be unthinkable – and likely even just as centralized as BSC – to store such amounts of data, the Solana architecture was designed in a way that only requires nodes to store the last two days of transactions. The rest is permanently committed to Arweave – immutable and able to be accessed at any time by any application.
Philosophical trust issues aside, being able to recall the full transaction history of Solana is the only thing that makes it possible build apps on top. NFT marketplaces, exchanges, and block explorers simply wouldn’t function without a memory longer than two days.
This heroic feat of archivism is achieved using what’s known as the SOLAR Bridge – a connection between Solana and Arweave that rapidly sends confirmed blocks to Arweave in parallel.
This ensures Solana transaction history is always available to be validated and leveraged in the ever-expanding ecosystem of applications.
But Solana is just one part of the story – when it comes to archiving blockchains, Arweave isn’t limited to just one source.
KYVE is everything’s layer 0
KYVE is an Arweave-based network which can archive and validate data of any kind. The problem it solves is one present with the current Solana/Arweave bridge – the problem of trust and centralization. With just one node sending Solana’s transaction history to Arweave, we have to be able to trust that node.
KYVE is different. Anyone can be a validator, and KYVE is already backed by a number of blockchains which will fund future archiving projects, including Avalanche, NEAR, Polkadot, Cosmos and Solana. KYVE’s testnet is currently actively archiving Avalanche data and plans to expand far beyond that, powered by a recent raise of $2.8m. The second round of funding indicates interest from chains like Celo, Mina, Aurora, Moonriver, Moonbeam and Octopus Network.
“It’s completely decentralized. So the way it works, is that you have a pool that has a job of archiving a data stream. The cool thing about that is that in that pool you have one uploader. All that uploader’s job is that it pulls down data from the blockchain and pushes it to Arweave. Then you have validators in the pool – it could be any number – we currently have 1,300. Validators look at all the data an uploader is archiving, pull it down for themselves, talk to the truth of source, and then do a simple hash comparison and vote. We can then be sure that the validators are correct because KYVE is backed by proof-of-stake, and minority voters get slashed” – John Letey, KYVE co-founder”
KYVE itself doesn’t actually store any data itself, only validates it – “this is a perfect example of why Arweave is a layer 0”, KYVE co-founder John Letey told us.
When mainnet launches, KYVE will be archiving new blocks from the pooled chains but also tracing data all the way back to the genesis block, ensuring every transaction from every partner blockchain is stored immutably and verified by staked nodes.
Storing chain data on Arweave permissionlessly exposes it via a GraphQL API to any app that wants to use it, plus it’s got a user-friendly naming scheme for tags that allow any query to reference any transaction across indexed chains.
KYVE’s plans for the future involve rolling out a new pool with a new blockchain partner every week, and examining how zero knowledge proofs could also be implemented.
The universal storage layer of the future
At its heart, Arweave is a database. But unlike any database ever created, it’s permanent, immutable, and accessible by anyone. Through unique incentive mechanisms, it guarantees storage and access. Through a unique consensus mechanism, it remains lightweight and fast. And, with the addition of bundles, it has guaranteed reliability and is easy for application developers to integrate.
The Solana/Arweave bridge was the first step. Now, KYVE will take it even further by providing a simple, trustless and validated way for any blockchain to store its entire history. Arweave is the blockchain world’s layer 0 for these reasons, now and in the future.
