In the present paper blockchain-based data storage systems are getting reviewed. A comparative description is given to different data storage systems that use blockchain technologies in order to store data. The authors review pros and cons of the selected blockchain-based data storage systems. As a result, the optimal data storage system is chosen and suggestions on how to improve the storage systems are given.
This work’s goal is to give a comparative characteristic of the data storage systems mentioned before to show their differences with pros and cons to give further suggestions on how these data storage systems could be improved.
In [
The blockchain trilemma offered by Vitalik Buterin states that it is inevitable that these parameters will be traded between each other. Author of the trilemma states that blockchain itself is decentralized and security is an essential part of it. However, it affects scalability and Bitcoin network is a good example of it since reducing latency to improve transaction throughput may result in weakening of security due to a higher probability of creating forks in the blockchain.
It totally makes sense, and it affects blockchain-based data storage systems as well. Main purpose of this research is to find out which of the most popular blockchain-based data storage systems is the best for use. These three parameters from the blockchain trilemma will be reviewed as well during this research.
Authors of [
However, using blockchain-based data storage systems gives more privacy and security it
doesn’t ensure that nothing bad happens to data. There still could be malicious nodes, so data
tampering and other malicious activites, such as stealing of a content under the copyright,
spreading of illegal content, etc may take place. Ways of protecting from it were researched in
detail in [
For example, these measures offered by authors may help in preventing data losses or unwanted data access if the following measures are taken: System immutability: once data gets into storage it cannot be neither removed nor changed Decentralization: data must be stored in multiple different locations with an independent access to data to prevent a single point failure Digital signatures: ensure non-repudation by using signatures based on asymmetric key encryption Blockchain hashing algorithm: ensures integrity of data Data audibility: each node must be able to perform and audit Scalability Easy tracking: user should be able to see his data at any moment
These methods seem completely reasonable and having them applied would mean having the security grade of a given data storage system close to the highest grade possible.
In this part the problems of local storages and general cloud services are being stated to explain why it makes sense to switch from older technologies everyone is used to the newer ones. Also, there will be explained how the problem of selection of the optimal data storage system is going to be solved. In the last subsection there will be given more information on selected datastorage systems with their pros and cons.
First thing first, we should investigate feasibility of switching from local data storages and
usual cloud storages to blockchain-based storages. There are some problems that cause
switching from “non-modern” data storage systems [
If you give it a thought, there are many problems which are if not solved, then are greatly
relieved in blockchain-based data storage systems [
Main purpose of this work is to determined differences between the most popular storage systems [9] that use blockchain to operate.
For this research has been decided to use linear additive convolution with normalizing factors as it gives precise results while still being simple.
Next step after deciding on using this method is to decide on parameters which values will play its role in selecting some blockchain-based data storage as an optimal one.
There are the parameters which are crucial when deciding if the blockchain-based data storage is worth using it (grades will be varying from 0 to 10 where 0 – worst result of a modern worl, and 10 – best result so far): 1. Cost – represents the cost of using and deployment of a specific data storage system. Mark is based on usage cost compared to other systems [10, 11]. 2. Speed – represents the speed of the data processing and transmission. Speed estimates are based on total bandwidth flow and transaction processing speed for each system. 3. Security and privacy – represent level of users’ data protection and how unlikely the data could be stolen or accessed by hackers. Evaluated relative to other blockchain-based data storage system in comparison. 4. Memory capacity – represents the amount of memory available or space for data storage in each system. The score is calculated relative to other systems. 5. Centralization – represents the centralization level of the particular data storage system. Even though it might count as a part of security parameter, we believe it should be a separate parameter as it is one of the main features provided by blockchain-based systems.
This method should let us compare all the chosen blockchain-based data storage systems to get a better understanding which one of them is optimal for migration there.
This part will serve as a guide to what each of the competing blockchain-based data storage systems is [12-14] and what is its pros and cons. It will reveal more details about these systems which should give a better understanding of the current state of each system. 1. BitTorrent – a protocol for sharing and distributing files which uses a decentralized network to upload and transfer parts of files between users. The only cons of this data storage system are connected with it being secure so nobody has access to uploaded files, therefore the data could contain a harmful software or illegal content (either protected by copyrights or just forbidden by law).
Main benefits of BitTorrent: Decentralization: users can upload and share files simultaneously without overloading centralized servers; Upload speed: sharding allows faster receiving of big amounts of data Bandwidth: each user uploads and receives files at the same time which lowers the pressure on the network TRON-based: affiliation with a famous and long-maintained network is a plus since this is more stable, so data should be accessible all the time and won’t get lost in the future due to the network’s shutdown 2. Filecoin – decentralized network which allows storing and sharing of data using cryptocurrencies and blockchain technologies.
Main pros of Filecoin: Decentralization: the network uses nodes all around the world to operate so no main server is necessary for it to work High level of customization: storage providers are free to set the conditions of their data providing, be it the cost of the space they provide, who can use their service or under which jurisdiction they want to work with their potential client Proof-of-storage protocol: the system uses a unique protocol which checks if the data is really stored on chain before the storage providers receive their fee Easy provider switching: users are free to change their storage provider since there is an implemented content addressing for files so user doesn’t have to download their files to reupload them to another provider’s storage if they decide to do so Main cons of Filecoin: Entry barrier: Filecoin uses its own token for transactions inside the system what forces users to spend additional time to trade the tokens Prices instability: as all the economy is built on the tokens which are not some stablecoins, the prices can vary from very high to very low 3. Sia – decentralized data storage platform which uses blockchain technologies to create a distributed storage network.
Main pros of Sia: High privacy: the network’s decentralization makes any third parties unable to access your files or prevent you from accessing them Easily interactable: Sia exposes endpoints which are easily understandable since they are using JSON. The endpoints give access to the renter-host protocol which allows massive renting operations Open source: all software is completely open source which allows users to contribute to the code and even build applications and businesses on top of Sia High throughput: Sia stores redundant file segments on nodes across the globe providing high throughput Main cons of Sia: Potentially dangerous files: due to absence of control over transferred files (which is usually one of pros), may happen a scenario where illegal or harmful files may be transferred over the network Dynamic pricing: due to instability of tokens which are used to purchase services on the network, a provider may earn money which wouldn’t cost anything on the next day 4. Storj – decentralized data storage platform which uses blockchain technologies and cryptography protocols to maintain data security and privacy.
Main pros of Storj: Quick data transfer: due to operations being performed simultaneously data transfer appears to happen quicker Security: every file is encrypted which prevents data from being either censored, or hacked Sustainability: as promoted by the company, when using this system, due to absence of data centers, “user’s carbon footprint is dramatically reduced” Stable cost: unlike usual pattern where the cost of services may vary due to token price volatility, Storj provides costs in real money what makes it clear how much user has to pay Main cons of Storj: Data recovery: if a storage of one of the nodes gets broken, then part of data may get lost Hosting complexity: process of becoming a storage provider is difficult due to users reviews
Summing up all the information given in this section, these systems look alike with most of their features being similar, yet with little details added to some of them.
In order to decide on optimal blockchain-based data storage system the earlier considered methods will be put to use.
First of all, we have to grade every system by selected criteria: BitTorrent: 1. Cost (7/10): offers good prices in comparison to cloud storages, but not cheapest in blockchain-based data storages industry 2. Speed (8/10): known for its pretty fast data transfer capability due to its peer-to-peer technology 3. Security and privacy (8/10): provides good level of security and privacy, but not the highest in the industry [15, 16] 4. Memory capacity (8/10): offers a large amount of storage, but not unlimited 5. Decentralization (7/10): is decentralized, however its level of centralization is lower than level of its competitors [17] Filecoin: 1. Cost (8/10): offers a competitive pricing which makes the model affordable for many users 2. Speed (7/10): its speed is decent, however not the transfers are not the fastest 3. Security and privacy (9/10): known for its good security measures [18] 4. Memory capacity (9/10): has a large decentralized storage network 5. Decentralization (8/10): is decentralized, but there are still some aspects of centralization Sia: 1. Cost (9/10): known for its high cost-effectiveness 2. Speed (8/10): offer a good speed of transfer 3. Security and privacy (9/10): provides a high level of security and privacy due to its decentralized nature 4. Memory capacity (8/10): offers large data storage, but not unlimited 5. Decentralization (9/10): the system is highly decentralized Storj: 1. Cost (8/10): offers competitive pricing model 2. Speed (7/10): good transfer speed, but not the best 3. Security and privacy (9/10): offers a high level of security and privacy due to its decentralized nature 4. Memory capacity (8/10): offers large data storage, but not unlimited 5. Decentralization (8/10): the system is decentralized, but there are still some aspects of decentralization
Below you can find Table 1 with all this data:
8 9 8 8 7 8 7 8 9 9 9 8 9 8 8 7 8 9 8 The formula (1) of the convolution with normalizing factors is as follows: ∑ , (1) where
are normalizing factors.
∑ Here are the normalizing factors for the convolution: Cost: 1/ (7+8+9+8) = 1/32 = 0.031 Speed: 1/ (8+7+8+7) = 1/30 = 0.33 Security and privacy = 1/ (8+9+9+9) = 1/35 = 0.29 Memory capacity = 1/ (8+9+8+8) = 1/33 = 0.3 Decentralization = 1/ (7+8+9+8) = 1/32 = 0.31 Next step is counting of the convolution index.
Results will be as follows: BitTorrent: 7/32 + 8/30 + 8/35 + 8/33 + 7/32= 1.175 Filecoin: 8/32 + 7/30 + 9/35 + 9/33 + 8/32 = 1.263 Sia: 9/32 + 8/30 + 9/35 + 8/33 + 9/32 = 1.329 Storj: 8/32 + 7/30 + 9/35 + 8/33 + 8/32 = 1.233
Now when we’ve got every data storage system’s convolution index the experiment is over and it’s time to sum up all our results.
In the Table 2 below there is data from Table 1 with respective convolution index in front of every present data storage system.
So, even though the results are very close, the Sia has the highest convolution index. The lowest one, on the other hand, was acquired by BitTorrent. The final results are with each blockchain-based data storage system’s placement is shown in Table 3 below:
So, according to the results, Sia received the highest grade, Filecoin has the second highest grade, Storj has the third one and BitTorrent has the lowest. The results do not differ very much. However, they allow us to tell which of the data storage systems is slightly better.
The method we used allowed us to choose optimal blockchain-based data storage system having in mind all the parameters are equally important to us [19, 20]. It allows us to spot even minor differences while deciding on which of the data storage systems is better.
First of all, it is important to discuss which other ways to discover the most suitable blockchainbased data storage system we had. There are some more ways that come to mind if we decide not to use the chosen method: Feature by feature comparison – rough method which doesn’t give any precise results and is strictly subjective. The point is to choose main features and compare data storage systems to one another. After finding differences, choose which one has better ones. Hasn’t been selected due to being inaccurate. Linear additive convolution with weighted coefficients – another type of convolution which is more suitable if we have exact preferences on features. Difference with the chosen method is that multiplicators are selected based on our preferences instead of sum of grades on the parameter. Hasn’t been chosen because it’s been decided to study generally better blockchain-based data storage, not raiting one parameter over another.
Summing up the analytics above, it seemed more logical for us to use linear additive convolution with normalizing factors.
Since we’ve got the results, we can now also form suggestions on improvement for every system on the list: BitTorrent: Since it has the lowest grades for almost every parameter, then it would have a lot of aspects to work on. The cost problem could be fixed by adding liquidity to the chain’s token. It would add some stability to the token price and storage costs should be reduced by the market. Regarding speed, memory capacity and decentralization, their situation could be improved by attracting new users into system, which, due to the blockchain trilemma, would lower the level of security [21]. The security level’s loss can also be compensated by adding the features from the research [22]. Filecoin: It has almost the highest marks of security and memory capacity. The security is again to be improved by adding measures from [22-24]. Speed, on the other hand, has the lowest value among all the participants. It can be improved by enlarging the quantity of storage providers or getting new storage providers with high storage capacity; however, it would lower the security level. Costs could be normalized by, again, adding liquidity or getting more users. New users would also raise level of decentralization and speed level in exchange for privacy [23, 24]. Sia: It has got the biggest grade for almost every parameter. However, it can still be improved. It is mostly about getting new customers and storage providers with bigger storages and in more places around the world. Storj: Since it has its cost in dollars, then liquidity suggestion would not help there. However, all other tips given to another data storage systems are applicable there.
As many can notice, the ways to improve blockchain-based data storage systems are not too diverse. Therefore, all the systems have received pretty much alike suggestions. In [25] authors propose a model for secured data storage in decentralized cloud by using Etherium. The proposed system solves the challenges of cloud computing by integrating MetaMask for user registration, Ethereum smart contracts for secure data storage, ECC for efficient encryption, and IPFS for decentralized file sharing. It provides enhanced security, user privacy and decentralization by leveraging the strengths of each technology in a unified approach. Alike methods are also mentioned in [26, 27].
The purposes of the research were to find optimal blockchain-based data storage system, to see how they differ and what can be done to improve them.
During the completion of this paper, we’ve researched differences between most popular blockchain-based data storage systems in order to decide which of them was better. It appeared that Sia was the better one, Filecoin was the second one, third one was Storj and Bittorent was the last. However, it doesn’t mean any of them is either bad or superior – they are all decent and have its own pros and cons. Then we have added suggestions for each one of them on how they can be improved.
The choice of method to decide which one of the systems was better has been highlighted as well as it was explained why other methods would be less effective during this research.
However, as stated before, this work doesn’t give a definite answer to the question “What blockchain-based data storage system is better?” or “Which blockchain-based data storage system should I be using?” as there are still a lot of subjective factors like looks of interface, companies-partners, etc. which have not been taken into account since it has been decided to avoid subjective factors during decision-making to make the decision more clear and objective.
However, we think this topic deserves more researches from other perspectives or using other methods. In example, it is worth to make some research with statistics gathering on the subjective aspects of the given blockchain-based data storage systems. It would give even more precise results than this research and could even probably answer the questions stated before.
The research study depicted in this paper is funded by the French National Research Agency (ANR), project ANR-19-CE23-0005 BI4people (Business intelligence for the people). [27] C. Zakaret, N. Peladarinos, V. Cheimaras, E. Tserepas, P. Papageorgas, M. Aillerie, D.
Piromalis, K. Agavanakis; Blockchain and Secure Element, a Hybrid Approach for Secure Energy Smart Meter Gateways, Sensors 2022, 22(24), 9664; https://doi.org/10.3390/s22249664.