In 2014, decentralized storage projects like Maidsafe, Storj, and Sia rolled out to provide data services with storage functions. The concept of blockchain-based decentralized data storage was first introduced at this time, much earlier than most of the recent popular decentralized projects. However, the development of this particular infrastructure requires substantially more intricate underlying technologies making progress slower than most DeFi applications.
Decentralized data storage is a critical component of the Web3.0 infrastructure and the success of Web3.0 itself is highly contingent upon. This article analyzes the progression of the commonly used protocols in this track from an observational perspective and the specific economic models caused by the protocols’ characteristics and limitations.
1.0 Generation for Decentralized Storage: Cryptocurrency as A Means of Circulation, Unrelated to Storage
Storj and Sia are two representatives of the 1.0 generation of decentralized storage. Their protocols are missing the on-chain logic commonly implemented in the latest iteration of storage projects. Simply put, the design of these two storage networks is similar to centralized storage; only the server location is decentralized with simple redundancy mechanisms added to ensure data security. Storj issued tokens centrally to incentivize miners while users pay for storage space and traffic in U.S. dollars. Although Sia’s tokens are also issued based on PoW mining unrelated to storage, its proof of storage contracts are being executed on-chain.
With minimal exposition, the notable characteristics of these projects are very apparent: the [project’s] tokens issuance is unrelated to storage specs and users simply use tokens to pay for storage. Sia has improved their iteration’s encryption, redundancy, and storage contracts features, but its economic model in which token issuance is unrelated to the project’s core storage business is fundamentally flawed.
Imperfections aside, being the two of the first storage projects on this track has allowed Storj and Sia to gain a vast client base. According to their official websites’ data, Sia currently has 1915 T.B. data storage and 6.5PB available storage space; Storj clocks in at 5.8PB storage space with an unknown data storage capacity, but 418 million documents on their website.
This large user base clearly points to the effectiveness of Storj and Sia’s storage technology. That being said, their economic model cannot guarantee the quality of the services provided by miners or the earning situation of miners. It cannot balance supply and demand, and token price entirely depends on market trends. In the long run, this economic model cannot sustain the ambitions of the project.
2.0 generation: Proof of Storage On-Chain Validation, Incentives are Related to Storage Specs
At the onset of storage 2.0, Filecoin achieved its iconic status. This heavyweight project has raised more than 100 million U.S. dollars through its ICO in half an hour. After several years of development, it was finally launched officially in October 2020. From my standpoint, the greatest value in Filecoin is its well-thought-out on-chain logic.
I will not go into the details to introduce Filecoin’s mechanism since relevant information can be found on the internet. Instead I will focus on analyzing some of the innovations of Filecoin’s protocols from the perspective of the project’s underlying logic: leveraging algorithms to improve service quality.
Motivating decentralized miners to provide high-quality services has always been the primary obstacle for storage protocols. Filecoin ensures miners’ service quality through algorithms including Proof of SpaceTime and Proof of Replication. It uses sophisticated algorithms to not only verify the availability of storage space, but also the validity of the data stored therein. Usually these statuses are verified by the service providers’ reputation in centralized storage.
It is a significant improvement that Filecoin can confirm storage services and decentralize the following storage activities, but the workflow is perplexing and consumes many network resources. Despite this, Filecoin is continuously improving and optimizing its protocols to cut down the resources required for on-chain proofs to reduce the overall price for building decentralized storage.
A Closed-Loop Economic Model
Tokens are the means by which users pay for storage services in the 1.0 decentralized storage generation, forming a fragile economic model which takes a long time to balance supply and demand. Filecoin has created an economic model with low staking to limit malicious mining and reign in the high-purchasing-demand in the secondary market. It provides a buffer zone for the project to develop soundly as user demands are not high in the early stage of the project.
Additionally, Filecoin empowers miners to generate blocks (including transactions and orders verification) through a weighting algorithm which is proportionally correlated to the size of the miner’s submitted sector space. This further helps incentivize miners to provide more storage for the network. While the model as a whole offers a fair distribution of incentives, it does cause other problems.
Limitations of Filecoin’s Protocols
Filecoin’s approach of using blockchain to verify the validity of storage has changed the underlying logic of decentralized storage and redefined the public storage chain. Unfortunately, the limitations of these improved protocols came to light not long after its implementation. Filecoin, as the bottom public chain, gave miners too much power thus allowing them to manipulate the network to maximize their rewards. This in turn undermined the project’s long-term benefits. This type of miners’ manipulation has happened to Bitcoin and Ethereum. Ethereum’s solution to this dilemma has been sharing its profits with miners. Currently, Filecoin’s miners are providing services and packing blocks–this is like athletes playing the role of referees within the same game. As a result, it is not uncommon for projects to get manipulated by some of the giant nodes.
2.5 Generation: Replacing Proofs with Incentives
Readers familiar with the storage track already know I am about to mention Arweave. Arweave is defined as 2.5 instead of 3.0 due to it being an underlying public chain, but Arweave’s mechanism has a fundamentally different innovation than that of Filecoin. Arweave directly stores its data on-chain and each block generation is based on the data of a previous random block using PoW. Miners then store as many blocks as possible to increase the possibility of generating a new block. This facilitates the generation of excess copies of data that can be stored permanently which we refer to as “permanent storage.”
Many Web3 projects have paid attention to Arweave this year, given that its protocols are the most logical choice for Web3. The ability to make multiple data backups without complex verification and its enhanced performance in interoperability with Web3 projects makes Arweave a true frontrunner. Nonetheless, the fundamental design of its mechanism and the flaws in its economic model caused Arweave’s inability to provide services for large-scale commercial applications.
In contrast, Filecoin allows on-chain applications to schedule storage resources, a crucial feature that Arweave lacks which causes slow network speeds. Large-scale businesses have ever-increasing demand for high-speed and efficient networks beyond simply the need of data storage. Another difference between these two projects is that Filecoin still has retrieval miners similar to the current CDN (content delivery network), whereas Arweave’s lack of an incentive mechanism to target this issue is hindering its performance notably. In spite of these limitations Arweave can still satisfy many Web3 applications with low data transmission and processing demand.
Let’s show some data to substantiate my point. First and foremost, the data storage volume on the Filecoin network is 1,500 times more than that of Arweave. Sia from the 1.0 generation has nearly 100 times the data storage volume of Arweave. Arweave’s protocols are logical for Web3, but the data and limitations of the project is showing that not all innovations are better than tradition.
Arweave’s one-time payment model also has many drawbacks as well. Miners continuously pay for expenses to provide services and the one-time payment structure makes miners likely to quit if expenses get too high. Since Arweave relies on miners taking the initiative to make enough copies to ensure data availability, any data held by an individual miner without a backup will be lost permanently should the miner quit. Even though the chance of this occurring is low, it forces us to reassess the potential risks of data security.
3.0 Generation: The Arrival of Decentralized “Cloud”
Using the infrastructure built by their predecessors, the core value of the decentralized “cloud” is finally incorporated by the storage projects that have rolled out this year. These projects, including CESS, Crust, Stratos, and others provide more comprehensive solutions overall for the majority of their practical applications. Let’s use CESS as an example to examine the implementation mechanism of decentralized “cloud” storage projects without getting into too many details about the on-chain proof algorithms.
The multi-layer network architecture in CESS includes storage, scheduling, consensus and the application layer that forms the entire network. Each model or layer has its incentives mechanism based on its functions to optimize the whole network’s performance. This aspect alone drastically increases network efficiency and stability, solving the slow network issue that Arweave faced due to its lack of incentivization of the important CDN. The layered structure provides a high degree of scalability too, as its functions are separated into different layers which divides miners’ responsibilities to avoid the dilemma that happened in Filecoin. CESS also separates its block generation and storage functions, but uses an algorithm to allocate partial block rewards to miners. This allows consensus miners to approve the confirmation of transactions and orders to achieve fairness within the overall network.
On top of all this, the previously mentioned multi-layer network permits the use of CESS for large-scale applications by dispersing its storage and CDN service providers to improve data transmission efficiency in the decentralized network.
CESS is a one-stop platform for complete data services, including but not limited to data preprocessing(copies, slicing, encryption, and redundancy) and an application layer complete with an API interface and a smart contract platform. Protocols with these functions embedded can provide complete solutions for users and save resources to the greatest extent possible. It is true that other storage protocols can develop privacy protection and other effective functions on their existing networks, but CESS’s ease of use associated with this full-featured network which supports EVM and WASM is second to none in terms of flexibility and convenience for developers.
Prospects of Decentralized Cloud Storage
The data infrastructure in the current 3.0 generation is constructed based on the previous effort of its predecessors. It continues to improve and fix the known issues of the previous generations. Current storage protocols were developed primarily in reference to Filecoin and traditional cloud storage protocols, replacing central authority with on-chain cryptographic algorithms for proof of storage. Theoretically, blockchain-based cloud services can be widely used if the processing of on-chain transactions reaches the same efficiency as these centralized protocols.
In my opinion, the payment methods permitted for any given storage fee is the crux of most current protocols. Using tokens as a means of payment could be the solution for Web3, but cost assessment requires a stable calculating method and is difficult to indoctrinate with the high fluctuations of token value. In order to design a practical economic model, current protocols should issue stable coins on-chain or make stable coins available cross-chain, and the protocol tokens should play the role of stock equity in a traditional business. In short, the protocol tokens measure the project value while the stable coins act as mediums of exchange for products and/or services within the project. As stable coins support more blockchains and continuously develop their cross-chain availability and programmable features, altogether, they will constitute an effective and viable economic model for decentralized storage.