A comprehensive comparison of Filecoin and Storj: Which is the better decentralized cloud storage?

Original title: "Who is the better decentralized cloud storage? Filecoin vs Storj design points and actual test PK"
Written by: First Class Warehouse

There are currently two main long-term views on Filecoin, the storage king (pigeon king) in the market: bearish and bullish, but many people are just simply talking about the bulls and bears. What's worse, there is no logic and they are exaggerating the long-term value of FIL - unfortunately, no complete logic has been seen so far.

Today, TouDengCang will conduct an in-depth discussion on the mechanism design of Filecoin, deeply analyze the business model, and conduct a simple test of the system, in order to demonstrate as detailed as possible all the basis and logic of our long-term research on Filecoin.

It’s time to pour cold water on the storage king Filecoin with reason and evidence.

Exploration of decentralized cloud storage

In the decentralized cloud storage sector, there are currently only four top-tier projects that can store data: Filecoin, Storj, Arweave , and SiaCoin.

SiaCoin is too centralized (there are only two mining pools, and one of them is closely related to the project owner, see the 13th issue of the First Class Warehouse Wealth Code for details), and its incentive mechanism has serious problems, so it is no longer worth paying attention to. Of the other three, Arweave actually set up a separate portal and built a permanent storage "Library of Alexandria". This path and ceiling need to be considered separately, and the valuation also needs to be calculated separately. Strictly speaking, Arweave and Filecoin are not suitable for horizontal comparison.

The only two heavyweight players that can truly be considered to be competing head-on in the decentralized cloud storage are Filecoin and Storj.

There is no doubt that they have chosen two completely different directions in the matter of decentralized cloud storage:

Filecoin is determined to achieve the perfect decentralization in the minds of blockchain geeks at all costs, regardless of storage performance and experience; while Storj, based on the implementation of node decentralization, focuses on providing high SLA service standards and high-stability performance to meet the high requirements of enterprise-level storage, and reduces high costs to meet the picky practical needs of the long-tail enterprise market.

Or to put it simply, one of them pursues better decentralization, and the other pursues better cloud storage, but have they really achieved it?

Both projects have been "working hard" for several years. TouDengCang has conducted many brainstorming sessions internally. Below, we will start a comprehensive comparison of Filecoin and Storj from the perspectives of decentralization and core storage business. The core of this article is divided into two parts. The first half compares Filecoin and Storj as comprehensively as possible through more than ten aspects such as decentralization, token model, incentive mechanism, mining machine threshold, product positioning, etc. The second half compares the storage business capabilities horizontally.

Since this article is quite long, with a total of more than 17,000 words, we have worked hard to type out every word. Of course, we hope that you can read the entire article, but we understand that you may not have the patience to read it all, so we directly list our conclusions in each title for your convenience. Among them, No. 1-9, 14, 15, and 16 are contents that you are strongly recommended to read carefully.

(In the actual test that lasted for several days, TouDieCang found that due to the major defects in Filecoin’s incentive mechanism, miners have a relatively high risk of centralization. This is elaborated in detail in points 2 and 16)

In-depth comparison between Storj and Filecoin in terms of decentralization, incentive mechanism, token model, product positioning, etc.

1. Product positioning: Filecoin has a grand vision, but the reality is far from it, while Storj has a smaller vision, but it has been well implemented so far**

Filecoin is mainly positioned as the incentive layer of IPFS to solve the current efficiency problem of IPFS. The core point is decentralization, and the main target group should be decentralized geeks and DAPPs who pursue data ownership freedom. We do not think that traditional storage needs will be transferred to the Filecoin network at this stage, because the current test results of Filecoin's storage capacity are really surprising.

Storj is positioned as a decentralized cloud storage platform that competes with Amazon S3, similar to Uber's role. It integrates and utilizes idle resources to output standardized products. Its main user group is all users of Amazon's S3 storage service. Storj provides the same SLA service standards, faster upload and download speeds, better privacy, decentralization, and the price is only half of Amazon's similar products.

In general, Filecoin has a more ambitious vision, but it has just been launched and its storage service capabilities are relatively poor. Storj took Amazon S3's service standards as a milestone at the beginning of its planning, and officially launched it after reaching the milestone. Therefore, Storj is obviously better in the commercial cloud storage part. The entire storage and access experience is closer to the storage needs of real users, with the ability to store and delete, fast speed, flexible supply, and you pay for the amount of service you use.

2. Network decentralization: Although each has its own advantages, Filecoin is better overall.

Storj has a very clear "triangular relationship": user end, satellite, and node. Storj's satellite is not a satellite flying in the sky, but a server cluster. Storj's storage node network currently has more than 8,000 nodes, mainly in Europe and the United States, with fewer nodes in Asia. Currently, there are more than 11,000 registered users and more than 3,000 active users. There are currently 4 satellite clusters, three of which are in commercial use and one is used as a test satellite.

The relationship between the three roles of Storj is as follows: the user side issues a demand and uploads and downloads data, the node is responsible for storing and downloading data to the user side, and the satellite assumes a series of important roles: selecting the fastest node for the user side when uploading and downloading, recording expenditures and income for both the user side and the node, recording the "location" of the user's stored files, and the "reputation data" of the storage node - you can think of the satellite as establishing a credit system similar to Alipay's credit score for the node to manage storage nodes. It is undeniable that the existence of satellites makes Storj relatively decentralized in terms of network structure. At the same time, satellites also have a single point of failure risk, that is, if the satellite is attacked and exploded, all the content stored through the satellite will be lost. However, the official explanation of Storj is that the satellite is not a single server, but a server cluster, which will not fail easily, and there will be plans to open satellites in the future so that anyone can build satellites.

Figure 1: The triangle structure of user terminal, satellite and node

In terms of storage, Filecoin has roles including storage users, storage miners, and retrieval miners. The role of retrieval miners is a bit similar to that of "satellites", but the difference is that retrieval miners are completely decentralized, and anyone can become a retrieval miner. Retrieval miners do not audit storage nodes like "satellites", and retrieval miners do not audit Filecoin's storage nodes. From the perspective of this network structure, we believe that Filecoin will be more decentralized.

Figure 2: The architecture diagram in the Filecoin white paper, showing the miner market and retrieval market structure

Figure 3: A simplified schematic diagram of Figure 2

The storage process of Filecoin is as follows: users can choose the miners they want from a group of storage miners to place an order and pay the order fee at the same time. The miners then accept the order, store it, and package the sectors. When retrieving, retrieval miners are automatically assigned. After the retrieval miners retrieve the data, they send it to the user. While sending the data, the user pays the retrieval fee.

Although we believe that Filecoin will be more decentralized in terms of architectural design, does Filecoin definitely win in terms of the degree of decentralization?

Things are not that simple. Although Filecoin does not need the role of satellite, it has very high requirements for mining machines, and ordinary computers cannot meet the standards. Therefore, Filecoin is destined to be impossible to achieve Storj, where you can just buy a 1T hard drive (about 200 yuan) and a 250 yuan Raspberry Pi to mine at home - the price of mining machines that meet Filecoin requirements is more than 100,000 yuan. Storj's low hardware threshold allowed the number of nodes to reach 100,000 when the V2 version came out. Currently, the V3 version has nearly 10,000 nodes. Moreover, we believe that there may be a long period of time in the future when the scale of Filecoin nodes will be one or even two orders of magnitude less than Storj. For example, now, Filecoin nodes: 600+, Storj nodes: 8000+. In terms of the number of nodes, Storj will be better.

In addition, from the perspective of node distribution, as shown in the figure below, Filecoin nodes are distributed all over the world in a relatively balanced manner, with many in Asia; while Storj is spread across Europe, North America, South America, and Australia, but relatively few in Asia, especially in China, where there are less than 100 nodes. This directly leads to the upload and download speeds in China being significantly lower than those in Europe and the United States. In this sense, we believe that Filecoin is superior in terms of node distribution range.

Figure 4: Storj's node distribution map. You can see that there are few nodes in Asia, with more than 8,000 nodes.

Figure 5: Filecoin miner node distribution diagram. As of 2020/10/27, the number of miners is 685.

Therefore, in this round of whether the network architecture is decentralized, both sides have their winners and losers. From the perspective of network structure, Filecoin obviously wins, while from the perspective of the number of nodes, Storj wins, and from the perspective of node distribution, Filecoin wins again.

But here we need to add a major hidden danger in what we think is the design of Filecoin mechanism:

Since packaging sectors requires large calculations, the threshold for Filecoin mining machines is extremely high, which determines that the number of Filecoin mining nodes will be limited, and most of them are concentrated in the data center computer room. Suppose a user has stored 100 movies, each movie is 10GB, and a total of 1TB of data is in a miner's hard drive. When 1,000 users want to watch these movies at the same time, the traffic generated may reach 1000MB/S, and the bandwidth of the miner will become a bottleneck, because high bandwidth also means high cost, and the download link, as a storage miner role to provide download bandwidth and traffic, this link income is zero, only the retrieval miner will charge a little retrieval fee, and according to the measured data, this retrieval fee is far lower than (one in thousands) bandwidth expenditure. Although Filecoin has designed to provide part of the block reward to the retrieval miner in the future, it is still on paper. We believe that the bandwidth bottleneck may become the Achilles' heel of Filecoin storage. In the future, due to the low packaging capacity of small miners (see points 4-7 for details), large miners will package files faster, so most of the user file storage will flow to large miners, which will also aggravate the outbreak of this bottleneck problem. We will continue to focus on observation.

3. Token model: Filecoin is the clear winner

We believe that Filecoin's token model, combined with the storage business, is a business subsidy model . The system issues tokens to subsidize the high mining machine costs of miners, so that miners can quickly recover their hardware costs through the tokens issued by the system (commonly known as mining rewards). Finally, investors in the secondary market take over to maintain high coin prices to pay for the costs. This is a typical strategy of paying for the cost. When miners recover their hardware costs through the token income from subsidies (mining), they no longer need to worry about hardware costs, so miners can provide cloud storage services to the outside world regardless of cost . This is the essence of the Filecoin economic model that Tou De Cang is thinking about. Such an economic model is very advantageous in terms of storage prices. This is like a famous beef offal store owner in Guangzhou who charged you five yuan for a large bowl of beef offal because he had 10 buildings to rent out during the demolition of his home and did not make money from the beef offal store. We believe that such an economic model has great advantages in providing cloud storage products that are far below market prices. In addition, it has also achieved good results in motivating miners and expanding the network scale, but in terms of mining mechanism design, it is obviously too harsh on miners, which is another matter (Detailed explanation in point 4). Overall, we believe that Filecoin's token mechanism design should be better than Storj.

There is not much design in the token mechanism design of Storj. It is to simply circulate Storj as the company's securities + currency. The economic model design of Storj is still based on the traditional concept of doing business. That is, a user deposits 1TB of data, which will be distributed to 80 miners for storage after erasure code redundancy, occupying a total of 2.7 TB of space. Then, the user is charged $10, and the miners are given $1.5/TB*2.7 TB=$4.05. Then, out of the remaining $4, $1 is given to partners and $3 is kept by the team. The remaining $2 is used for marketing activities or miscellaneous expenses. This is a very traditional economic model design. There is a weak relationship between business growth and token price increase. It can be said that Storj's token model design is not closely integrated with the business.

As a project established at the same time as Ethereum, the Storj project token model and economic model were designed in the early stage of blockchain 2.0 development. Now it seems to be old school (entering middle age). Although it is sincere, it is not sexy. First Class Warehouse has also made suggestions to the team, but the team obviously has no plans to change the model at present.

In this round, no matter how many problems there are with the Filecoin token model, Filecoin wins.

4. Miner incentive mechanism: Filecoin is harsh on miners, but miners can make money, Storj is gentle to miners, but miners cannot make money

Storj's miners (nodes) correspond to the storage miners in Filecoin.

Regarding the miner incentive mechanism, we understand its goal as: to encourage miners to stay online for a long time and stably, so as to provide stable storage services. Based on this goal, Filecoin and Storj also use the pledge method to ensure, but Filecoin uses pre-pledge , that is, miners must buy coins to pledge before working, and then get paid after finishing the work. Storj uses post-pledge , that is, part of the miners' salary income is pledged. The former is more ruthless to miners, while the latter is more gentle to miners. The following is the details.

Storj's incentives for miners include several parts:

1. A node reputation system requires miners to prove that they can be online for a long time and provide high-performance services at any time (more than 95% of the time online) + miners must prove that they store data + prove that they keep data intact at any time (no loss or damage), so that they can or have the opportunity to earn income. This process can take several months. Miners with poor performance will be eliminated, and qualified miners will earn income, which includes content storage fees + upload and download traffic fees;

2. Nodes with higher configuration and faster response speed will be given priority and can obtain more data and higher income;

3. From the 1st to the 9th month after joining the mining, part of the miner's income will be pledged in the system. Miners should try to keep operating for at least 15 months to get all the income. If the miner quits before 15 months, or is disconnected at random after 15 months, or is "fired" by the satellite due to poor performance, then part of the income will be deducted, which is equivalent to using the previous income as a pledge. Having said so much, in fact, after the author's actual test, basically as long as the miner is not disconnected every day, it is not a big problem to simply be a miner, and you don't have to worry about the equipment.

Through such a mechanism, Storj is obviously incentivizing miners to do their best to operate for at least 15 months, greatly reducing the loss rate of miners. The lower the node loss rate, the higher the durability of customer files (less likely to be lost). Of course, it doesn’t matter if some nodes suddenly fall off. Each file has 80 fragments distributed around the world. Even if a war tsunami occurs somewhere and a batch of nodes are lost, the remaining nodes will automatically repair other fragments to keep the files available. Currently, the Storj network stores 170 million files, and the number of file losses is 0.

The storage miner incentive mechanism of Filecoin is not much different from that of Storj, although the technical paths used are different:

1. Through the system, miners still need to prove that they have stored the data (Proof of Replication PpRep) + prove every 30 seconds that the data they stored is still there and complete, not lost or damaged (Proof of Space and Time Post), waiting for users to retrieve files at any time, but miners do not need to accumulate credit for several months like Storj miners.

2. If a miner is offline for a short period of time or the storage sector fails, a fine must be paid on a daily basis. If the miner forgets to declare the failure, a fine must also be paid. If the failure lasts for more than 14 days, the sector will be shut down and the miner will be fined 90 days of block rewards. Only by fully operating the promised order cycle can the miner get back all the FIL pledged previously.

Filecoin encourages miners to stay online in this way, which is the same as Storj's original intention, and the punishment is not light. It can motivate miners to stay online to a certain extent.

The different parts of the miner incentives are mainly about pledge, income rules and mining machine subsidies: the current effective computing power of the Filecoin network is 660PiB, and the FIL pledged is 13.6 million. On average, for every 1TiB of computing power, an average of 20.1 FIL is pledged. According to the latest data statistics, for every 1TiB of computing power added, an average of 11.18 FIL is pledged. The reward distribution mechanism is block reward + storage fee. The block reward is 30% simple casting + 70% benchmark casting. The block reward will increase with the increase of the effective adoption rate of the network. The current block reward is 11.37FIL/block, but it needs to be distributed linearly to miners within 180 days (daily income is averaged to the next 180 days) to encourage miners to operate for a long time. At present, according to the FIP-004 proposal update, in order to reduce the pressure of miners' pledge, the mining reward is changed to release 25% of the block reward first, and the remaining 75% still needs to be distributed linearly to miners within 180 days.

At the same time, since the system cannot currently distinguish between the data generated by miners themselves and the actual user data, Filecoin has designed several rules, such as verified users and committed capacity, to incentivize miners to store real data as much as possible.

Verified users refer to users who have been verified to have real data needs. Storing such user data will bring storage miners 10 times the income of other data. The details of verified users are described in detail in the "9th point: authenticity of stored data volume" section of this article.

In terms of miner incentives, Storj's incentive mechanism is simple and clear, and this set of rules can well serve the goal of high performance and high reliability of its system; and although Filecoin's rules have made great efforts to serve decentralization and have achieved decentralization in architecture, there are no rules that can effectively prevent miners from engaging in negative or malicious behaviors such as brushing data and refusing storage orders, nor can they incentivize miners to provide good enough cloud storage services. Miners are profit-seeking, and the design of the mechanism makes miners tend to provide storage space (sector packaging) rather than storage services (miners can refuse orders, miners do not need to guarantee bandwidth, and miners are not responsible for transmission speeds) . From this point of view, although both have achieved the purpose of mechanism design to a certain extent, the storage service provided by Storj will be more stable, and the upload and download test in point 16 later in the article also verifies this problem. However, although Filecoin is harsh on miners and has a strict punishment mechanism, FIL is more profitable. The scale effect allows miners to invest a lot and produce a lot. Storj pursues the decentralization of the network, so the official limits the scale of miners to fit the growth rate of storage demand. Therefore, Storj miners cannot invest on a large scale, and can only invest individually. For example, if a person puts a NAS at home, it can only store 30TB of data at most. If you add a NAS, it will be identified as the same node in the Storj network, so data will only be allocated to you according to one node. This means that miners cannot invest on a large scale, and at the same time, they cannot have large-scale benefits. So from the perspective of miners' income, Filecoin is ruthless, but it can make money; Storj is gentle, but it can only make a little money and cannot make a fortune.

5. Mining machine threshold: Storj is much lower than Filecoin

The threshold for Storj's mining machines is relatively low. The minimum requirements are: one processor core dedicated to one node, one 500G hard drive, 2TB of broadband traffic per month, at least 5MB/S upstream and at least 25MB/S downstream. With such a low threshold, almost every household can install their computer into a Storj node. The lowest cost is 800 yuan to assemble a NAS. I have installed several nodes for testing.

The threshold for Filecoin mining machines is much higher: the official basic equipment requirements for mining nodes: AMD 8+ core CPU, at least 128GB of RAM (memory), a powerful GPU (the average number of cores is 3000, currently generally using 2080Ti/3080/3090 graphics cards), a solid-state drive larger than 512GB (select 1TB SSD NMVe interface). Currently, the price of a single Filecoin mining machine in Chengdu is over 100,000.

Obviously, the threshold for Filecoin mining machines is much higher than that of Storj. From an economic perspective, this will obviously result in fewer people being able to buy mining machines to work as miners. The direct result is that the number of nodes will inevitably be small. We expect the number of Filecoin nodes to be 1-2 orders of magnitude less than that of Storj. The current data shows that the number of Filecoin nodes is 600+, and Storj is 8000+. Fewer nodes will make Filecoin far weaker than Storj in terms of traffic diversion and CDN service provision. Once Filecoin storage applications are rolled out, Filecoin's large miners will also face greater traffic pressure.

6. File retrieval and deletion: Storj wins

Storj generates a "metadata" for each file uploaded to the network, which is the address of the file storage location. This address is stored in the satellite. Based on this address, the satellite can retrieve the file. This process is very fast and simple. At the same time, it can also perform operations such as deletion, transfer, and copy. According to the actual test of the first-class warehouse, it only takes 3 seconds to delete a file from Storj. Storj's search fee is a very small number ($0.0000022/time), which can be ignored unless the search volume is huge.

Filecoin's file retrieval mainly relies on the retrieval miners in the network. The retrieval miners publish retrieval requirements on gossipsub based on the file information provided by the user, then obtain the files from the storage miners and send part of the files and pricing requirements to the user. Finally, the user confirms the order and pays the fee, which is relatively convenient for users.

Since Filecoin packages and encapsulates files, each sector is folded layer by layer into a huge data crystal. Deleting data from it is equivalent to dismantling the deeply folded warehouse layer by layer and finding an object to throw out, which is very difficult - it takes 40-50 minutes to package the sector. Therefore, the files in Filecoin cannot be deleted by users at present. If you want to delete a file, it can only be done through negotiation between the user and the storage miner. The first-class warehouse consulted the official, and the official reply about file deletion is that the file can only be deleted after the order of the file expires. At the same time, the storage time of Filecoin starts at 180 days, because the minimum life cycle of a sector is 180 days, that is, as long as you store data in Filecoin, it must be stored for at least 180 days, and the user must pay the storage fee for 180 days in one lump sum.

In addition, after uploading a file, Filecoin needs to wait until the miner successfully packages the file into a sector before it can be retrieved and downloaded. The packaging time mainly depends on the hardware capabilities of the miner. If a small miner is slow, it will take several days, and if a large miner is fast, it will take an hour or so. Currently, the first warehouse has been tested and the fastest is to upload it at night and package it into a sector the next morning. Moreover, when Filecoin retrieves (downloads) files, the delay is quite severe and the network is unstable. It usually takes at least ten seconds, and sometimes three minutes, before it starts downloading. When Storj downloads files, it starts downloading directly and is not sensitive to delays.

In addition, because Filecoin is a public chain, you must submit an order to the miners every time you upload or download, and you will pay a fee when submitting the order. Therefore, to successfully pay, you must ensure that the block is synchronized to the latest block. Therefore, under the current circumstances, to upload or download something to the Filecoin network, you must first synchronize dozens of GB of block data, and the block data is still growing at a rate of 400M per day. This is a fatal flaw that currently affects the popularization of Filecoin applications. Therefore, Filecoin officials have also proposed plans to develop light nodes that do not require synchronization of the entire blockchain, which should be able to solve this problem. As an ERC20 token, Storj does not have similar problems.

From the above rules and actual test results, it is clear that Storj is closer to the actual storage user needs than Filecoin in object storage service, with storage and access at any time, fast retrieval, flexible fee settlement, and the amount of space and traffic used will be deducted. In contrast, Filecoin depends on whether the block synchronization is completed, and the fee settlement is an unfair clause. It charges 180 days of fees without saying anything, and the stored data cannot be deleted. In general, we believe that Storj wins in this aspect.

7. File encryption and privacy: Encryption by default vs. No encryption by default, Storj wins

Storj uses client-side encryption. All data that needs to be stored in the network is encrypted by the client before being uploaded. The encrypted data needs to be decrypted with a key before it can be viewed. The key is stored in the local computer where the client is located and cannot be obtained from the network. In other words, even if the user's data can be found through the satellite, it is impossible to know what is inside. Moreover, the file is divided into 80 copies and distributed to the hosts of 80 miners. It is necessary to obtain the data of at least 29 miners at the same time to get the complete file. After getting the complete file, it is necessary to use the key to decrypt it, which basically eliminates the possibility of data leakage.

The data that Filecoin needs to upload is not encrypted by default, which means that it is up to the user to decide whether to encrypt it. Users have the possibility of uploading unencrypted data. In addition, Filecoin is a single miner for a single user, or multiple miners with multiple copies for a single user. The user's data is directly exposed to the miners, and the file's CID can be obtained at will as long as it is known by others. The privacy is relatively poor. It can be said that the mechanism design of Filecoin does not consider too many privacy needs, and only allows users to encrypt files by themselves.

Therefore, we believe that Storj is the clear winner in terms of privacy.

8. Data maintenance: Storj is better

The highlight of Storj's data maintenance is the use of erasure coding technology. When each file is uploaded, the data will be redundant by 2.7 times, and 80 fragments will be generated and stored in 80 nodes. Any 29 mining pieces can generate a complete file. The satellite will check from time to time to determine whether these 80 fragments are still there, and the number of them damaged and lost over time. Once the number of existing fragments drops below 50, the satellite will start data repair, obtain 29 fragments from existing nodes to regenerate the file, and then slice the lost fragments and re-store them in other nodes to ensure that the network has enough fragments. The whole process is imperceptible to users.

Filecoin data is maintained by miners, and the role of maintenance miners is also played by other storage miners. Users store files on multiple miners. If the data of one miner is damaged, other miners with the same data will transfer and repair the file. (If the user only stores one copy, there will be no maintenance miner, and if it is lost, it is lost)

In terms of data maintenance, there is no doubt that Storj wins with its thorough and perfect design. The Filecoin mechanism relies entirely on users to choose the data redundancy multiples and to punish miners under extremely strict conditions to ensure file durability. Due to the strict punishment mechanism, miners will indeed actively maintain the security of files, but natural disasters and man-made disasters are inevitable. Even Alipay can be mined by forklifts, let alone these miners concentrated in data centers? At present, the number of Storj file losses is 0, and Filecoin has just been launched. The actual effect still needs time to test, but from the perspective of user experience and mechanism design alone, we believe that Storj will be better.

9. Authenticity of stored data volume: Storj slightly wins

If we look at this issue from the perspective of results, it is somewhat unfair, because Storj was launched half a year earlier than Filecoin, and currently has 8PB of real business data stored on it, while Filecoin has only been online for half a month, and it is not even a full month. So when we discuss this issue, we will not look at the results, but mainly discuss it from the perspective of mechanism.

In terms of mechanism, Storj is a decentralized storage. When users upload files, they do not know which nodes the files will be stored on (and they are most likely distributed all over the world), so they cannot store them for themselves. Moreover, the direct cost of uploading files is very different from the income. Users have to pay $10/month to upload 1TB, while miners only earn $1.5/month for storing 1TB. Therefore, miners have no motivation to upload false files.

But Filecoin is not like this. Storage users can directly designate miners to store their own files, and the cost of storing 1TB of files is less than $3 per month. If miners encapsulate this 1TB of files into sectors, they can mine 6.6 FIL per month, and the current price is $21.7. In addition, in order to encourage miners to store "valid" data, the official proposed the concept of "verified users". To put it bluntly, the storage users are certified by community voting. The data stored by the verified users is recognized as "valid" data by the network, and miners can get 10 times the computing power by encapsulating these "valid" data. In this way, encapsulating 1TB of data can get $210 per month. Such a mechanism will definitely make miners strive to pretend to be a "verified user" who needs a lot of storage. As long as they pass the community certification, they can get a large amount of quota, and then designate miners to store it for themselves, which is perfect.

The following is a brief introduction to the community governance "verified user" community certification "opening" process:

The entire governance is divided into three roles: root key holder, notary, and customer group. These roles can be anyone. The root key holder is selected by the community, and his main task is to hold the root key and unconditionally execute community resolutions. The notary is also selected by the community, and his main responsibility is to allocate capacity quotas to customers who apply for quotas, and to apply for more quotas from the community. After the community votes, the root key holder will execute the quota activation. After the quota is activated, it can be allocated to customers for "effective" storage.

10. Theoretical vulnerability of the network (whether it is easy to be attacked): Storj is better

Filecoin miners are vulnerable to attacks. Miners are prone to DDOS attacks that cause bandwidth to be occupied. If miners are unable to submit storage proofs to the network in time due to attacks, they will be punished by deducting 90 days of block rewards and staked tokens after more than 14 days. The storage sector is considered an invalid sector, and the computing power allocated to the sector is reset to zero. At this time, miners can choose to discard the file, which means a high risk of data loss for users. During the testnet stage, there was a case where miners had their computing power reset to zero due to attacks.

This attack is not a problem for Storj miners. Any operation of the Storj client on the node must be performed through the satellite. Even if individual nodes are attacked, it will not affect the retrieval of user data, unless more than 51 of the 80 nodes storing the user data are attacked at the same time, resulting in the data being unable to be retrieved and repaired. This difficulty is obviously much more difficult than attacking a single Filecoin miner. The biggest risk of Storj lies not in the nodes, but in the satellite. However, the satellite is not a single server, but a server cluster. The cluster's anti-attack ability is obviously stronger than a single server.

And because Filecoin's user storage data is stored by a single miner, once the miner network is delayed or attacked by traffic, it will be difficult to retrieve the user data. Therefore, for useful data, users still have to choose multiple miners to store it.

Storj's data is stored in fragments, so as long as 29 of the 80 nodes are online, the file can be retrieved completely. Relatively speaking, the reliability is much higher than Filecoin.

11. Community support: Juan, you should hire more community managers.

This is a point I really don’t want to mention, but Filecoin is doing really poorly in this regard, so I would like to take this opportunity to draw official attention.

Filecoin's main community is Slack. When you consult Slack about current mechanism issues, system bugs, update issues, or miners' mining questions, you will get a slow response, which is basically concentrated at one time point, and sometimes no one even responds. According to the response situation, it seems that there is only one community leader who mainly responds to questions in the community. Most of the time, it is the same person in the chat room who comes out to answer questions. However, some questions are not officially confirmed, so we dare not believe them casually! I want to say that Filecoin has also raised a lot of money. I heard that there are hundreds of people in the development team alone. Why don't you recruit a few more people to respond to community questions properly? This has invisibly discouraged many people who want to understand Filecoin. I can't understand it myself, and I can't ask the community.

In addition to regular municipal meetings every quarter, Storj has Twitter, forums, and Support platforms to submit questions. The first-class warehouse encountered some problems during the testing of the Storj network. When consulting the official website on these platforms, almost all questions will be answered the next day. Judging from the reply situation, several people are responding to questions from various different fields, basically being able to answer questions at all, and being friendly to various details, and are willing to disclose most of the data.

12. Financial transparency: This two sides are almost the same

Storj will announce token spending and locking situations according to quarterly. The chain can also check its token locking situation, but the whereabouts of its fiat currency fundraising funds are not announced. It is a relatively transparent project. According to the current official quarterly financial report, the coins in the team are enough for him to spend more than ten years.

The Filecoin blog has not announced the use of funds, with a total ICO of US$257 million. We are not sure how much funds the team has. If anyone has relevant data, please provide it.

13. Market Value & Price Expectations: This is not important

In fact, we don’t want to explain this part too much, but if we briefly talk about the result, it is suspected that we will call someone for orders, so let’s explain it briefly.

Storj's current price is around 0.35 yuan, 312 "special price" this year is 0.08 yuan, with a short-term high of 0.73 yuan, with a total current total of 425 million yuan, and the official holds 262 million yuan, and the market circulating 162 million yuan. In the future, as the business volume grows, the tokens in the official hand will be gradually released into the hands of miners. Only when the tokens in the official hand are released can it be changed to buying coins from the market to pay miners' storage fees. The release speed is proportional to the business volume. The more business volume, the faster the release, the more circulation. The official holds coins is currently a relatively prominent issue affecting the price of Storj. If there are major changes in this part in the future (such as destruction), then with the current growth rate of Storj's business volume, there are still opportunities in the future.

For Filecoin, it can only fall all the way. At present, FIL can maintain a high price because of the design of the pledge mechanism, which forces miners to take over the market to increase computing power. But this is definitely unhealthy. There is no sufficient, free market game, and no sufficient chip exchange. It is unscientific to rely solely on pledge and lock positions to lock in liquidity to lock in high prices and high market value. This is unscientific. If FIL maintains the overall market value of tens of billions of dollars, there will be no investment opportunities. Just watch the show with the leeks. However, as the mining coins continue to be released and the entire circulation continues to increase, we believe that the price will still slowly go down. It will eventually reach a balance, that is, the willingness of miners to buy coins and the balance between selling and cashing out by coin holders. The key can be paid attention to the comparison of token release speed and pledge speed. After the lending market is launched, you need to pay attention to the interest rate and borrowing volume of the lending market. In the long run, we are still optimistic that Filecoin will have the opportunity to solve some problems in the existing storage field , but if Filecoin today has a market value of 60 billion yuan and Filecoin five years later has a market value of 120 billion yuan, what is the significance for you and me? Isn’t buying big cakes delicious?

PK of core business: Who can provide better cloud storage

To answer the question of who can provide better cloud storage, we first need to understand the service standard SLA of cloud storage, which is currently the gold standard for commercial cloud storage in the world. We use the current SLA service standard of Amazon S3 cloud storage as the standard line to measure the performance of both parties in the core cloud storage business (but we will not list the specific content of Amazon SLA. If you are interested, you can go to Amazon's official viewing). Amazon S3 SLA mainly includes several dimensions:

14. Storage Durability and Availability: Storj is better

Storj mainly uses erasure coding technology to solve the problems of durability and availability. Erasure coding technology mainly divides files into 80 equal parts at 2.67 times redundancy, and distributes them to 80 nodes on the network. Only 29 of them are required to be online to restore the complete file. Through this mechanism, Storj achieves 99.99999% file durability, currently stores 170 million files, and the number of lost files is 0. Because Storj V3 network design was designed to benchmark the SLA service standards of Amazon S3 cloud storage at the beginning, Storj will be relatively better in this part of its capabilities.

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

15. Storage pricing and mechanism: Filecoin is definitely cheaper

Storj's pricing is to benchmark the service price of Amazon S3 storage, and combined with its network's main components as idle hardware resources, and due to the erasure coding, Storj will automatically have 2.7 times redundancy of stored content. In summary, Storj's storage service is priced at half of Amazon. The data storage service price of 1TB is US$10/month, and the data traffic fee is US$45/TB. The data paid by Storj to miners is US$1.5/month, and the traffic fee is US$20/TB. Since Storj's main goal is To B users, a centralized pricing method is adopted to price in fiat currency to avoid token price fluctuations and provide a unified market price.

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

In addition, compared with the two models, Storj focuses on using idle resources and turning waste into treasure, providing high SLA-standard cloud storage products through erasure codes and storage node management rules. Storj still provides cloud storage services with a more traditional business logic. The business logic is: suppose Storj purchases 2,700TB of capacity per month from storage miners at a cost of US$1.5/TB, providing it to users with storage needs, and selling 1,000TB at a price of US$10/TB (the erasure code generates 2.7 times redundancy), the interest rate spread in the middle is the money earned by the team. We believe that the method of low-price wholesale, selling at a price of US$1.5/TB, and making interest rate spreads makes Storj look very traditional.

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :

16. Storage upload download speed: I have to say, Storj wins

Mechanically speaking, Storj uses erasure coding technology. The file download process is to select 35 fastest nodes from 80 storage nodes for competition transmission. In the end, the client retrieves data fragments from the 29 fastest nodes. Storj is fast, first of all, multi-point transmission is faster than single-point transmission. One problem of centralized cloud storage is that files are transmitted from a single point to the client during download. The transmission speed is very limited by bandwidth. Storj uses multi-point transmission to distribute bandwidth requirements to solve the problem of single-point bandwidth limitation - the bandwidth of 29 nodes is shared by the transmission traffic. The Filecoin mechanism allows users to select a single or multiple nodes to store, but when actually downloading, it is still downloaded from a single node, so its speed will be limited by the bandwidth of a single node.

The second is node competition. Due to the existence of node competition, Storj's nodes are "forced" to maintain high bandwidth to ensure their high upload and download speed. Every time a user uploads data, 130 nodes compete for storage data fragments, of which 80 nodes that complete storage faster can obtain this part of the data and subsequent revenue, and the other 50 cannot get anything. When users download data, there will also be some slower nodes that get nothing. Therefore, the faster the upload and download speed of nodes, the more they can guarantee that they can obtain the opportunity to store data fragments/transmit data fragments and obtain income. This ensures the overall upload and download speed of the network.

The following is the first-class warehouse. In order to obtain the real upload and download speed of Filecoin and Storj, we conduct actual tests on its network. First of all, we must emphasize that we only need to understand the speed of both parties. We do not require the accuracy of the experiment. The test process is divided into two parts. One part is a test conducted in China, mainly on home bandwidth, and the second part is a test conducted by renting foreign servers and mainly on foreign network conditions. In order to let everyone better feel the differences in storage capabilities of the two networks, we should restore every detail of the test process as much as possible, and also leave behind information such as CID. You can verify it in the block browser . The entire test process is as follows:

Local testing

The author is located in Xiamen, China (because the first-class warehouse is in Xiamen), uses a home 200Mb telecom broadband, and the test file size is 1006MB.

Filecoin Testing

After importing Lotus, generate the file fingerprint CID:
bafykbzaceb7jcjsnozpv45wpqyqrzilmehqgzin5ozhfwfvfbu7rn7oon3iu2

Several miners with suitable conditions were randomly selected from Slate.host

Select miner f09675 for the first time and submit storage order (Deal)

Generate the order CID in 30 seconds:
bafyreiheedhw7veoryrux6pznhrpagl32v5sb26gelhcpozwqwgits3dwq

But about 6 minutes later, it was found that Deal disappeared. The possible statement was that miner f09675 refused to trade.

So the first submission failed .

The second time selects miner f08157 to submit storage order (Deal)

From submitting the storage order (Deal) to the final StorageDealTransferring (file transfer is completed to the miner), it takes 17 minutes and 19 seconds. The specific steps are as follows:

Step 1: Submit the storage order (Deal), and the time it takes 1 minute and 14 seconds to generate the Deal CID. The generated order CID (DEAL CID):
bafyreidk5lzgtagrromd3qs4rk7akd57keaz7jlj5bibu7jhiy3wguf6sm

Step 2: The order status is converted to StorageDealClientFunding, which means that the miner is checking the balance, checking the order, etc. It took 4 minutes and 48 seconds

Step 3: File transfer, which takes 11 minutes and 14 seconds

Step 4: Generated Deal id: 961927, Deal status is displayed as StorageDealSealing (in encapsulation)

If you only take the file transfer time in the third step to calculate the file upload speed, the average speed should be 1.49MB/S

If the time from Deal is taken to the completion of file transfer, the average speed is 0.97MB/S

Which speed you think is better for representing the upload speed of this test depends on your choice. I have to remind you that the mechanism of Filecoin single point (user) transmission to single point (miner) determines that the speed is not the Filecoin network is good or not, but the network connection between the user and the miner they choose is smooth and high-speed. Therefore, this test can only represent the author's own network situation and the network connection between the miner he randomly selected, and it cannot represent whether the Filecoin network is good or not. Everyone's test speed may be different. The information learned from the official Slack forum is that some people are as fast as more than ten MB, while others are as slow as 250KB. Everyone should look at it rationally.

The rest is to wait for the miners to encapsulate it into the sector, with the expected time being several hours to several days, and there will be no statistics during this period.

(I have been waiting for two days and still not encapsulated into sectors.)

After waiting for two days, miner f08157 has not yet successfully encapsulated the file, so it is impossible to directly test the download speed of the miner node, so it has to download the same file from other miners who have successfully encapsulated for testing. Currently, miner f01782 has been selected to complete the encapsulation, as follows to retrieve (download) the file from the miner's node.

The first time I ran the retrieve command, I only successfully retrieved 1.05MB of data, and then got stuck for 10 minutes, but failed to retrieve it

The second time I ran the retrieve command, it showed Recv: 0 B, it was stuck for more than ten minutes, but the retrieve failed

The third time I ran the retrieve command, but the result was the same as above. It was stuck and the retrieve failed. I learned about the reason. It turned out that the block of my local Lotus was stuck in synchronization and it was not synchronized to the latest block.

After a whole night, the block height was still not synchronized, and the block synchronization speed was significantly lower than the new block generation speed. From this point of view, the block synchronization problem of the test network was still not well solved. It was clearly synchronized yesterday, and there were also many domestic nodes. Why is the synchronization speed so slow today, and it is impossible to download the file as long as the block is not synchronized. I went to the official forum to find a solution, and the result was to change the computer and use AMD's CPU + NVMe solid-state drive (I am so poor, I can't do it). Finally, I had no choice but to reinstall it all. After a few hours, I finally synchronized to the latest block and successfully conducted the fourth download test.

The fourth time I ran the retrieve command, I stuck in the DealStatusCreatingChannel for 3 minutes and 13 seconds. I thought it might be a wall blocking it, so I tried to turn on the VPN for 20 seconds, then turned off the VPN, and then went to the next step smoothly and started to transfer files. The file transfer process took a total of 3 minutes and 28 seconds, with an average speed of 4.83MB/S. In order to make me look more rigorous, I downloaded and tested it again. The download process took 9 minutes and 11 seconds, with an average speed of 1.82MB/S . In order to save some face for Filecoin, I should take 4.83MB/S as the result of this download test. At this point, I experienced three failures and the fourth success. The download test was finally completed.

Storj speed test (under equivalent hardware and network conditions)

The upload test of Storj is relatively simple. I chose the Asian satellite https://asia-east-1.tardigrade.io/. After registering above, I got the API and then downloaded Filezilla. Using Filezilla, you can directly connect to Storj's Tardigrade network, which is also relatively simple.

The upload process is to configure Filezilla, select the file to be uploaded, and click upload directly. The upload is a file of the same size as Filecoin test. The upload time is 14 minutes and 27 seconds, with an average speed of 1.16MB/S . However, during the upload process, every 64MB upload of the file will pause for about 1 minute. The main reason is that Storj needs to cut the file into 64MB segments, and then redundant erasure and divide it into 80 small portions. These times slow down Storj's upload speed. Therefore, it took more than 14 minutes.

In addition, due to the fact that there are too few Asian nodes, Storj's upload is not stable when uploading large files in China. I uploaded them 5 times before success. Most of them are stuck when uploading to 60%-80%, and then the feedback failed to timeout.

The download process is relatively simple. Select the file directly and click to download. The download took 2 minutes and 21 seconds, with an average speed of 7.13MB/S , and the download was relatively stable. After trying several times, it was successfully downloaded directly.

Foreign server testing

The foreign server uses Vultr's most equipped Bare Metal, which is geographically located in Singapore, with a 4-core 8-line CPU and 32GB of memory. The file is still the same file, with a size of 1006MB and the file CID remains unchanged.

Filecoin Testing

The first time a suitable miner was randomly selected from Slate.host and f020904 to submit a storage order (Deal)

Generate the Deal CID in about 30 seconds:
bafyreicnzkb7csxxqwaqsbzxieoktlsk7ro42ad6slgncfoqmkmsmwg2ui

After about 4 minutes and 30 seconds, the order disappeared inexplicably after the order status changed to StorageDealFundsEnsured. As it turned out, the order submission failed!

The second random selection of miner f09569

Step 1: Deal CID was generated in about 28 seconds

Step 2: Just flash by, too fast and I didn't see clearly. I just think it's only one second

Step 3: Deal enters StorageDealTransferring state and starts transmitting data, but this link uses 32 minutes and 15 seconds (shock)

Step 4: After the file transfer is completed, the Deal id is generated: 962551 Deal status is displayed as StorageDealSealing (in encapsulation)

Based on the 32 minutes of file transfer in this test, the file upload speed is 0.52MB/S . I feel that I may have encountered a hot miner, or I have not had enough relationship with this miner. In order not to wrongly accuse Filecoin, I decided to choose a miner again and test it again.

The third time I really couldn't stand it anymore, so I decided to choose a well-known miner. I used the query-ask command to inquire for the top five miners in computing power. Only f01782 can get the price. The other miners did not respond when inquiring, whether they were using domestic clients or foreign clients. Didn't they open the door to welcome user data? As a Filecoin miner, if your hard drive is just for mining, what's the difference between you and BHD? Let's despise it first!

Step 1: Submit a storage order (Deal)

Generate the order CID after 27 seconds:
bafyreiczimrrwlhv2v2csn7bqmnq7yivtq4st4i3eyybkpiun3esxejnhy

Step 2: Just flash by, too fast and I didn't see clearly. I just think it's only one second

Step 3: Deal enters StorageDealTransferring state and starts transmitting data. The entire link is used for 2 minutes and 31 seconds.

Step 4: After the file transfer is completed, the Deal id is generated: 962675 The order status is displayed as StorageDealSealing (in package)

(After getting up the next day, the Deal status is already StorageDealActive, and the packaging is successful)

Calculated from the third test results, if the upload time is calculated in 2 minutes and 31 seconds, the average upload speed is 6.66MB/S. If the upload time is calculated from the time of submitting the Deal, the duration is 2 minutes and 59 seconds, and the average upload speed is 5.62MB/S.

Download the file by running the retrieve command to retrieve the file from miner f01782

After the first file retrieval, only 10MB of file was retrieved, an error timeout was reported, and the retrieval failed

The second file retrieval takes 3 minutes and 53 seconds. Average speed is 4.32MB/S

The author downloaded another 1.03GB file from the node of miner f020904. It was uploaded to test it before. It took 1 minute and 34 seconds to download the file, and the average speed was 11.21MB/S . I think this should be one of the problems with Filecoin, which is that the storage performance and bandwidth provided by different miners are inconsistent. There are big differences.

Storj speed test (under equivalent hardware and network conditions)

Using a foreign server to upload files to Storj, the upload speed is obviously much faster. I still chose https://asia-east-1.tardigrade.io/ Asian satellites and directly configure uplink to the server. Start the speed test:

1. Upload test: Select the same file, size 1006MB, upload time 58 seconds, average speed 17.34MB/S

2. Download test: Download the same file to the server, size 1006MB, download time 1 minute 45 seconds, average speed 9.58MB/S

The final inductive test results are as follows:

Domestic environment:

Filecoin upload speed: 1.49MB/S, download speed: 4.83MB/S

Storj upload speed: 1.16MB/S, download speed: 7.13MB/S

Foreign environment:

Filecoin upload speed: 6.66MB/S, download speed: 4.32MB/S

Storj upload speed: 17.34MB/S, download speed: 9.58MB/S

: : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : With the current storage characteristics of Filecoin, we cannot imagine how developers can use Filecoin network as the storage layer**, such as selecting miners for file storage; the storage order is limited to at least 180 days; the content cannot be deleted after uploading; anyone can get files, etc.

After the full in-depth PK is completed, we will look back and think about the topic. Who is the better decentralized cloud storage platform?

We believe that this issue should be divided into two parts and is more appropriate. The first part is who can be more decentralized, and the second part is who is a better cloud storage platform.

Regarding the question of who is more decentralized, the first-class warehouse is that in terms of mechanism design and network architecture, Filecoin is more decentralized, while in terms of node number and file fragmented and decentralized storage erasure coding rules design, Storj is more decentralized. If we have to choose one of the two, then we think Filecoin will be better.

As for who is the better cloud storage platform, we believe that Storj has done much more than the current Filecoin in cloud storage. For the purpose of providing high-speed and stable storage services, I think the advantages and disadvantages of Filecoin and Storj are very obvious. Because Storj has a high SLA service standard, the entire network performance unified, high durability and high availability guaranteed by the erasure and coding mechanism, the upload and download speed is only affected by the user's local network conditions, which is more conducive to storing useful data (I dare to store it myself). Filecoin is currently in the unified performance and service standards of miners, which is flawed in this matter. For B-side users, if there is important data that needs to be included in Filecoin, it is really necessary to choose a good Chinese miner. One is not enough, and it may take three.

Finally, as mentioned in points 2 and 16 above, due to the defects in Filecoin's economic model and miner incentive mechanism design, there is a great possibility for miners to centralize in the future: data storage and download traffic are likely to be collected on the mining machines of large miners in the future, and miners' export bandwidth will become a bottleneck in the entire network. It is hard for us to imagine that such Filecoin can serve as a storage layer to carry future decentralized storage of ZB orders of magnitude (1 billion TB). Filecoin needs to undergo major disruptive mechanism reforms, or even multiple transformations like Ethereum, to truly carry its mission. We will focus on the development direction of Filecoin, and will push it to members of the first-class warehouse when there are major changes in the mechanism. If the Filecoin mechanism does not undergo major innovations, then we wish you happiness!