Decentralized Cloud Comparison: Filecoin & Archon Cloud
As demand for data and sentiment against centralization grows, decentralized cloud storage solutions are cast in the spotlight. Decentralized cloud storage projects are gaining ground in the minds of developers who want to build applications and protocols free of the centralized control of tech giants who currently rule large portions of the web. Filecoin made a big initial splash in 2017 with their record breaking 257M ICO. This article compares Filecoin with a promising decentralized cloud upstart from San Francisco called Archon.
At a high level, both projects implement a blockchain-secured incentive layer with the ability to match storage buyers and storage sellers, but there are many key differences between the two projects. Filecoin has a head start, a track record with IPFS, and lots of development. Archon touts some major technological advantages and its focus on blockchain protocols and dapps gives it a unique product-market fit. Will this be enough to differentiate Archon, and be where frustrated Filecoin users go when they are fed up with its tech flaws?
The following table highlights some major differences (green means better).
This article compares the two projects in four interesting categories:
- infrastructure,
- upload/download process,
- user experience, and
- business/progress.
Infrastructure & Protocol
Both projects are developing their own native blockchain with native tokens. They are both specifically designed for file storage, are designed to scale, and designed to be decentralized. Filecoin’s blockchain implementation is more concretely detailed, and Archon has not yet announced specific details for its blockchain. Overall, Archon’s protocol design is more efficient, which theoretically enables faster DApps that can use more decentralized data. Archon has fewer performance limitations than Filecoin because it can scale faster from its more efficient cryptographic proofs, as well as its superior erasure codes.
Filecoin has a major scalability bottleneck due to its file sealing speed from its Proof of Replication. Both Archon and Filecoin use cryptographically secure proofs to confirm that storage providers are properly storing files, with Proof of Replication being one of them. This is significant because it allows the uploaders to be sure that their files are available. In Filecoin, this Proof of Replication maxes out at ~1GB/hour file sealing speed (as of Dec 2018), that means that a node can only make about 1GB of storage useful to serve in one hour, and this limits how much each node can store per annum and lowers profit. Archon has similar Proofs, but uses Provable Data Possession, which is very quick, so larger capacity hard drives can make more money in a shorter amount of time.
Both Archon and Filecoin can improve scalability if they introduce secure off-chain scaling solutions or on-chain improvements to increase Transactions-Per-Second (TPS), and thus increase storage capacity. Both can potentially scale to high TPS. Archon’s blockchain has high native TPS and low latency to maximize file uploads/second on the blockchain end, with downloads not bottlenecked by the blockchain. Filecoin’s blockchain seems to have a file transfer and storage limit, preventing it from scaling within a few orders of magnitude of traditional cloud like Amazon [1]. Many factors such as block-size and block-generation-speed limit storage capacity. Without additional changes, Archon also has the same issue.
Both systems also allow for encryption and file sharding, which can be adjusted based on parameters. Filecoin also focuses on file replication in addition to file sharding through erasure codes, whereas Archon focuses primarily on file sharding through erasure codes. Archon has the advantage in terms of file sharding due to its more advanced, proprietary set of erasure coding algorithms, which are used to create an erasure coded file format (the .af format) to allow far more efficient processing.
The Upload and Download Processes
Archon and Filecoin have quite different upload and download processes. This is important because the processes greatly impact the overall end user experience as well as the developer experience. Overall, Archon yields slightly better upload performance than Filecoin, and significantly faster downloads, especially for smaller files.
To upload a file on Archon, a file is sharded and distributed to a random set of storage providers fitting some matching/market criteria. This set of storage providers is selected by Archon’s blockchain-based routing. This is similar to the DNS that the internet uses in that it allows the download client to look up a file and decide which nodes to download from, but is specifically for selecting storage providers. The uploader and downloader both use this routing mechanism as a reference and establish connections with storage providers directly. These same nodes who store the files are responsible for providing bandwidth. Archon allows for downloads to be prepaid, so download time to first byte is comparable to centralized speeds and significantly faster than other decentralized storage solutions, including Filecoin.
Marketplace
The choice of how buyers and sellers are matched is dictated by marketplace design, and this can have a great impact on how efficient data can be transferred. Although Archon has a single marketplace for storage, in Filecoin, there are two separate marketplaces: the storage and the retrieval marketplace. The storage marketplace matches storage providers and storage buyers similar to how exchanges match buyers and sellers. For both projects, this process is done on-chain, and so the initiation for file uploads takes a while to register. After the storage provider and buyer are matched, they can exchange the data offchain. For Filecoin, however, the retrieval marketplace is an offchain orderbook that is gossiped to peers. This offchain order book relies on offchain payment channels, which (even if done extremely efficiently) causes the time to first byte to lag behind centralized solutions, so this means that Filecoin’s download speeds are slower as a result of its marketplace design.
End User Experience
Although both are seemingly targeted towards developers, Archon’s end product is more intuitive to use for both developers, taking design cues from existing centralized products. Archon has focused on providing good user experience from the start in addition to its tech and backend, whereas Filecoin seems to be solely focused on developing its tech and backend.
Since Archon is designed to be similar to AWS, it allows for uploaders to pay on the downloader’s behalf and for downloaders to download a file without any installation or knowledge of blockchain/tokens. This process works without downloading plugins, or confirming transactions, and this improvement to the experience may be what drives further traction.
Although Filecoin’s final product has not yet been released, if Filecoin’s usability is like that of IPFS, made by the same team, it may be very difficult to use Filecoin without prior development experience. According to technical user tests [2], Filecoin’s CLI is difficult to understand, and downloading a small testnet blockchain took many hours.
Archon also has crosschain integrations that benefits users in various public chains such as Ethereum, NEO, etc. Due to Archon Integration Nodes, Archon integrates directly with public chains to allow users in that ecosystem to pay with the native token (ETH in Ethereum, GAS in NEO), and also de-silos file storage from different blockchains. This means that a file can be potentially uploaded by a user in the Ethereum blockchain and be used by an EOS dapp. Filecoin has not stated any plans to integrate with other blockchains directly, but Filecoin may have easier access to existing Cloudflare users due to their partnerships.
Business & Progress
Filecoin has more progress than Archon and has a head start in timing and fundraising. Archon initially targets decentralized communities, public chains and dapps to drive adoption. Adoption across both projects is unclear at this point, but Filecoin’s strategy is difficult to guess from the lack of public information.
Filecoin’s first whitepaper came out in July 2014 and expects to launch the Mainnet around Q3 2019. Filecoin has a very active community on Github with a high amount of content, largely due to its pre-existing IPFS community as well as the publicity from their ICO. Its 252M raise gives them a strong war chest to recruit lots of talent. However, some investors worry over the relative secrecy and silence about its development progress until recently. There also seem to be many kinks and bugs to work out before it can develop a mainnet within a few more months. Filecoin’s testnet currently has >100 nodes globally [3], whereas Archon has less than 10 in its beta network; neither number is sufficient, where the target number of nodes is in the thousands or millions.
Archon started development in mid 2018. Archon is currently in public beta as of March 2019, with a mainnet scheduled for Q3 2020. Archon already integrated with one public chain, NEO, through its Archon Integration Nodes, which means Archon benefits from users and DApps already on NEO. It has also been rapidly expanded its network of partners within the decentralized ecosystem starting in May 2019.
Conclusion
Filecoin has obtained a large following due to its impressive team and big raise amount. However, Archon has advantages over Filecoin, but must grow quickly in order to challenge Filecoin for market share. Both are exciting projects in blockchain-based file storage, and are worth keeping an eye on as they grow.
References
[2] https://gergely.imreh.net/blog/2019/02/first-impressions-of-filecoin/
[3] https://stats.kittyhawk.wtf/
[4]https://medium.com/@sebastijaneder/decentralized-cloud-storage-comparison-storj-0chain-4fb751e45743
