Behind the rise of the distributed storage track: Filecoin's computing power is far ahead, and Arweave's storage capacity has increased 20 times in a year

This article is from PANews, authored by Jiang Haibo.

After three years of development, Filecoin experienced an initial price bubble upon launch, followed by a value trough at the beginning of this year, and now is undergoing price discovery. Recently, AR, STORJ, and CRU, which are also part of the storage sector, have reached new highs. This article will compare the usage of different projects, analyze the reasons for the rise in the storage sector, and observe how their growth has transformed from quantitative to qualitative changes.

Filecoin

IPFS is an open-source decentralized network protocol that allows users to store and transfer verifiable content-addressable data in a peer-to-peer network. While IPFS is an excellent tool for data storage, it is merely a network protocol, and it is challenging for ordinary users to become nodes and retain data. Moreover, IPFS itself lacks an incentive mechanism to encourage nodes to provide services for other users.

Thus, Filecoin was born. Filecoin serves as both the storage layer and the incentive layer for IPFS, and the nodes of Filecoin are also nodes of IPFS. The project incentivizes data storage miners and retrieval miners to maintain the network through FIL tokens, and customers pay for the services they require using FIL. Customers reach agreements with multiple nodes, which promise to retain a segment of data for a certain period; after the expiration, a renewal is necessary to maintain data persistence. For the same content, the longer the data is stored, the more FIL is paid, and data storage miners are not obligated to retain content indefinitely.

Filecoin uses Proof of Replication (PoR) and Proofs of Spacetime (PoSt) to prove that storage miners have retained a segment of data. Proof of Replication allows storage miners to prove that data has been stored on the corresponding device, while Proofs of Spacetime are used to check whether the storage provider has retained the corresponding data over a certain period. Filecoin nodes must encapsulate and store customer data and require a significant amount of cryptographic algorithms to ensure the security and effective storage of data. Therefore, ordinary home computers are difficult to meet the mining requirements of Filecoin. In addition to basic storage space and bandwidth, typical Filecoin mining machines also require:

• 8-core CPU;
• 128G RAM;
• Powerful GPU;
• 1T NVMe disk;
• Pre-staking of FIL tokens.

As of April 8, there are 1,885 active miners in Filecoin, providing 4.02 EiB of effective computing power. The largest miner offers 101.66 PiB of storage space, accounting for only 2.47%, indicating a relatively even distribution of storage capacity.

In the past 30 days, the total network computing power has increased from 2.09 PiB to 3.99 PB, a growth of 29.1%.

With the increase in computing power, the daily FIL reward for 1 TiB has decreased from 0.105 FIL to 0.0885 FIL over the past month, a decline of 15.7%.

Arweave

Unlike Filecoin's pay-per-time model, Arweave's biggest feature is the ability to permanently store data. Users only need to pay storage fees when adding data to the blockchain. Arweave believes that the cost of data storage will gradually decrease, similar to Moore's Law, and the interest generated from overpayment will help cover future storage costs, making a one-time payment sufficient to cover the cost of permanent storage. To ensure that data is permanently stored on the blockchain, some type of preservation mechanism is required. For example, on Ethereum, when running a node, the entire chain must be considered, and new data is added to the end of the chain, which continues to grow, representing blockchain-based persistence. Permanent storage also brings some issues, as the chain may become too large to maintain and store all data effectively.

Considering that all data is stored in blocks, as data grows, a single node may not be able to store the complete block data. Therefore, Arweave introduces the RadomX algorithm in mining to assess the completeness of blocks. The more blocks a node stores, the more likely it is to gain the right to package blocks. The project also incentivizes miners to permanently store historical blocks through Proof of Access (POA), rewarding miners not only for new block rewards but also for storing random old blocks. Miners are not required to store all block data, but the more blocks they store, the more likely they are to receive block rewards, and the more rewards they can obtain.

The hardware requirements for Arweave mining are close to those of a home computer. Besides basic networking, it only requires a 512G hard drive and 4-core 8G RAM, and can be deployed with a one-click solution from Ankr. The usage of Arweave is experiencing a qualitative change, with the current data stored in the network being 7.85 TB, compared to only 3.97 TB at the beginning of this year and just 378.15 GB a year ago, representing a growth of 1976% over the past year.

Currently, there are a total of 151 nodes in the network, mainly concentrated in Europe and the United States, with 69 in the U.S., 14 in Germany, 11 in France, and 9 in China.

The number of transactions has also reached new highs this year, with an average of 26,110 transactions per day over the past five days, compared to only 7,382 transactions a year ago, an increase of 254%.

The usage of data is also on the rise, with a single day's usage reaching 888 GB on March 26. The average usage for the previous week of this month was 13.33 GB, compared to 2.33 GB a year ago, an increase of 473%.

Storj

Storj is a distributed storage protocol based on Ethereum, developed by Storj Labs, with Ethereum founder Vitalik being a core author of the Storj v2 white paper. Storj aims to allow nodes to provide idle storage space and bandwidth, enabling other users to obtain distributed storage services at a better price. According to official comparisons, storing and downloading 1 TB of data using Tardigrade developed by Storj Labs costs only $55, while Amazon S3, Google Cloud, and Azure charge $108, $146, and $105, respectively.

Files are encrypted on the client side before being uploaded to Storj to ensure secure data storage. The encrypted files are divided into smaller fragments and distributed across various nodes. Data blocks are scattered across 80 nodes, and as long as 29 nodes are online, users can download the data. If a node wants to act maliciously and obtain stored data privately, it must collaborate with more than 20 other nodes, which incurs high costs, thus ensuring data privacy and security. The hardware requirements to become a storage node are not high, including:

• 1 processor core
• At least 550 GB of available disk space
• At least 2 TB of available bandwidth per month
• Minimum upload bandwidth of 5 Mbps
• Minimum download bandwidth of 25 Mbps
• Online 24/7.

With the recent increase in Ethereum gas fees, the transaction cost has risen to $44, making network usage expensive. With the help of Matter Labs, Storj now supports payments via zkSync, with funds held by smart contracts on Ethereum, and computation and storage conducted off-chain. Transactions are not verified individually but are aggregated for batch processing, increasing TPS and reducing transaction costs.

As of February this year, the Storj network has been operating on 11,000 storage node operators across 84 countries, with Storj satellites processing 160,000 transactions.

Crust

Crust provides a distributed storage network for the Web 3.0 ecosystem, supporting various storage layer protocols, including IPFS. The storage interface is open to application layer programs, allowing developers to build cloud storage applications and peer-to-peer content delivery networks (CDN) using the Crust/IPFS network. Crust builds a distributed ecosystem of privacy and ownership through Trusted Execution Environment (TEE) technology. Sensitive data is processed in a trusted execution environment to achieve hardware security. The Guaranteed Proof of Stake (GPoS) consensus mechanism allows more users to participate in maintaining network security, enabling users to stake tokens on high-quality nodes in the network to earn rewards. As an independent node, the requirement for a trusted execution environment necessitates hardware support for Intel SGX (Software Guard Extension). The required software/hardware conditions are as follows:

• Processor: Intel i5 7th generation or higher, 6 cores or more, must support Intel SGX;
• Motherboard: BIOS must support SGX;
• RAM: 32G or more;
• Operating System: Ubuntu 16.04/18.04/20.04.

Crust's Maxwell preview network storage market officially opened on February 28 of this year. As of April 8, there are 2,371 nodes in the network, with a total storage capacity of 330 PB, of which 2.74 PB has been used. Compared to the data in the March 22 weekly report, the number of nodes has increased by 50%, and total storage capacity has increased by 101%.

Conclusion

The purpose of distributed storage is to provide a secure, private, and cost-effective decentralized storage solution. Filecoin, Arweave, and Crust can all serve as incentive layers for IPFS, while Storj is built on Ethereum and can use zkSync L2 solutions for payments. Developed by Protocol Labs, Filecoin has a well-established mechanism, with computing power far exceeding other projects, and the total network computing power is still rapidly growing, having increased by 29.8% in the past 30 days.

The equipment required for Filecoin mining is the most complex, while Arweave and Storj can mine using home desktop computers. Arweave's storage volume has increased nearly 20 times over the past year. Because data in a distributed network is not subject to censorship, there may be illegal information stored in the network, such as anti-government statements contained in Arweave, which are visible to everyone, hindering mainstream adoption of the network.