Decentralized Data Storage — The Future of Data Storage?

Blockchain technology promises the world of decentralization. Demand for more freedom from centralized system is rising. Financial and payments system were among the first to be “blockchained”.

As an emerging technology, it has many flaws. It is expected that more research will be put around infrastructure before decentralized apps (dApps) and programs could run on the blockchain.

Now, the decentralized data storage will have the spotlight. It is become more and more of a crowded space as more projects turned into their heads, with Storj, Sia, Filecoin, Bluzelle and the recent newcomer, Arweave.

Decentralized storage is based on the idea of allowing people to rent out unused storage of their hard drive in a peer-to-peer network. With more different innovation taken by each team, this space will be an interesting topic to discuss.

Basics of Decentralized Storage

Decentralized storage works by spreading files across peer-to-peer network.

How does it work?

First, the file being uploaded results in the uploader retains the key to unlock the encryption. Then those files gets broken into several smaller pieces. Each of these small encrypted pieces gets duplicated guarantee redundancy. These files then sent to various computer among the peer-to-peer network.

Those computers holding the files only have a part of the file and it’s being encrypted. This means that those computer can’t have any information from files they are holding.

To recall a file, the uploader uses the private key he retains in the first place along with a blockchain-hosted hash table to locate all the pieces of the original file and order the network to restore the file.

Demand for Storage

Cloud storage market has the potential to grow to US$92.49 billion by 2022. Currently this market is dominated by centralized major players such as Amazon and Google.

With the premise of decentralized world, it could be a major turning point for customer to turn their heads to decentralized storage.

Competition among decentralized storage space would be tight given how large this market could become in the future.

Storj

Storj is by far the oldest project in this space with 20,000 users using its platform and 19,000 storage providers renting out their storage.

Storj architecture follows standard model of sharding, encryption and swarming as described before to split and rebuilt the files.

One of the advantage that Storj have is the user experiences they offer. Storj offers easy and intuitive platform to upload your files or to operate a node. It is clearly targeting mass market adoption with consumer without background or knowledge in blockchain could participate in the network.

They have pay-as-you-go payment model for uploaders and the good thing is they allow payment in USD. But, as you see later in this article, this payment model isn’t the best out there.

Storj doesn’t support mining. Instead, they utilized their storage providers or farmers by running them into file audits. These audits ensure that the farmer have an online connection to the network and holds a copy of the files. When all of that is confirmed, these farmers received payouts in form of STORJ tokens.

Storj have its own drawback as they aren’t fully decentralized due to the level of user experience that they offer. Although Storj is still more decentralized compare to current centralized solution and Storj also don’t have access to your files, you do have to trust Storj portal to run the network.

Another drawback from Storj is that they use Ethereum blockchain as they don’t have their own blockchain. With the current speed of Ethereum is disastrously slow, it’s a major threat for mass market adoption.

Sia

Sia is one of Storj’s competitor and it works in a similar fashion to Storj. To be exact Sia splits files into thirty segments, with enough redundancy that the could be recompile the file from only ten of the thirty segments.

Sia has its own blockchain, which is an advantage. With this proprietary blockchain, Sia supports smart contracts that able to manage and send files. Sia implements similar storage proofs to audit its nodes as Storj also have.

Compare to Storj, Sia doesn’t support fiat payments. Uploaders and storage providers get paid in Siacoin. Sia also needs miners to support its blockchain and those miners also being paid in Siacoin.

Filecoin

Filecoin is a project created by Protocol Labs. Protocol Labs itself already deployed other network-based storage, most notably IPFS.

Filecoin have massive backing from prominent Silicon Valley VCs such as Andreessen Horowitz. They also launched one of the most successful ICOs ever, raising $205 million in its token sale. While for me it’s not a good sign when blockchain-based projects raising massive amount of funding, it is a great achievement after all.

Based on its white paper, Filecoin architecture uses two types of network nodes. The first ones are storage nodes, the same as those in Sia and Storj. It is need to be noted that Filecoin seems to be approaching larger players with large amount of storage in offer. That’s not a good thing for sure.

The second ones are retrieval nodes. These nodes are physically surrounded by many storage nodes. They need high-bandwith, low-latency internet connection to retrieve and relay file shards in an efficient manner.

Filecoin works on proof of continuous replication of the file shards entrusted to you. Creating redundancy of files earns you Filecoin token.

Because it comes from the same creator of IPFS, Filecoin already have their own built-in protocol for redundancy and erasure.

Bluzelle

Bluzelle is kinda different than the other projects I mentioned above. While other projects such as Storj focused on storing files, Bluzelle aiming at developers and businesses that needs databases. It describe itself as “Oracle” in the decentralized system.

Bluzelle itself works by implementing swarming technologies in order to ensure developers having the highest throughput in performance, reliability and scalability. A swarm is a large group of nodes that work together to store and manage data.

As Bluzelle cannot be compared to projects above, I will only talk about how it compared to current centralized solution. Current centralized system met major hurdle in terms of scalability. This is where Bluzelle excels with the advantage of being decentralized, Bluzelle simply adds user to the network by incentives rather than building new data center that centralized system had to do.

Other than that, Bluzelle offers better privacy as nodes also cannot view customers data stored on the network. Also, the data stored in the network is immutable to breaches thanks to blockchain technology.

Aerweave

Aerweave is relatively new I this space as it is just finished its token sale. Although it is a newcomer, Arweave offers different and more likeable approach compare to usual monthly fees charged by data storage services — pay for storage one time only.

Depends on the lifetime of the storage, the cost needed to use the Arweave platform could be cheaper than recurring monthly fees from Storj, Sia or Filecoin.

Arweave itself is supported by its own blockchain structure called blockweave, designed to enable scalable storage in a cost efficient manner. As the amount of data stored in the system increases, the amount of hashing needed for consensus decreases, thus reducing the cost to store data.

Things that I do notice though is, Arweave somewhat lacking in their team experience as their team mainly only have academical background.

Wrap This Up

As I already talked about decentralized storage related projects above, I believe that the potentials are massive. With the competition being so fierce, it will end up resulting in great innovation among this space.

While these project have a long way to go, they will end up having face-to-face competition with Google or Amazon in the near future.