Filecoin, which has come out after a lot of calls, has been questioned as “centralized”. Will forks be a bitter medicine?
Written by: Luff
“IPFS wants to achieve data privacy and permanent storage, but Filecoin is gradually deviating from this original intention.”
At the just-concluded Xiamen World Mining Conference, MIX Chairman Han Weiping expressed the aspirations of the Filecoin community that was awaited.
Since the industrial revolution started the accelerator of human civilization evolution, discussions on technology and humanities have never stopped. Today, when giants dominate the Internet, people’s concerns about data privacy have led to reforms one after another. Correction through legal and other humanistic means is obviously a temporary solution but not a permanent cure. Distributed storage, which is expected to eliminate the possibility of evil from the technology itself, has begun to emerge temporarily, causing widespread concern and discussion.
Filecoin is the most watched practitioner of the “technological revolution” of distributed storage. However, three years after the launch of Filecoin, it has not only carried people’s expectations for distributed storage networks, but also suffered from the community. Questions about team centralization.
Filecoin was born: an incentive network built on the IPFS protocol
Let’s start from the earliest story of Filecoin and clarify step by step how it became a distributed network that has high expectations from digital nomads.
In May 2014, Juan Benet of Stanford University released a draft white paper describing the idea of a “peer-to-peer distributed file system”, which was the earliest prototype of the IPFS protocol. IPFS is a point-to-point hypermedia transmission protocol based on content addressing. It aims to improve the low transmission efficiency and single point of failure of the http protocol that has dominated the Internet for many years.
In order to better implement the idea of distributed file transfer protocol, Juan Benet established Protocol Labs in early 2015, focusing on the research and development of IPFS, and applied to join the prestigious startup incubator Y Combinator. A year later, Protocol Labs created four technical modules libp2p, IPLD, multiformats, and Orbit in one go. IPFS has achieved amazing growth. By June 2017, the IPFS protocol had stored 5 billion files. At this time, the blockchain industry is advancing with leaps and bounds, and the decentralized nature of the blockchain naturally fits with IPFS, and the two hit it off.
The early team of Protocol Labs, from Twitter
Soon after, Protocol Labs released the Filecoin white paper, and an incentive network based on blockchain technology built on the IPFS protocol was born. Filecoin relied on the aura of IPFS and Protocol Labs to raise a huge amount of US$257 million in the subsequent ICO, setting a record for ICO financing that year.
Filecoin is a distributed content storage and distribution network. It uses a proof-of-work mechanism similar to Bitcoin to incentivize network miners to provide users with data storage and retrieval services to obtain native token incentives, thereby achieving an incentive-compatible economic system on the IPFS protocol.
I have to admit that with strong engineering capabilities and abundant funds, the beautiful picture outlined by Filecoin seems to be in sight. However, there is inevitably a gap in reality, and Filecoin, which has a great start, does not seem to be smooth. Due to the wrong estimation of the difficulty of the project, the Filecoin main network was delayed frequently and gradually fell into silence. Until this year, the iconic progress of the testnet and the space race took turns on stage, and Filecoin once again returned to the public spotlight. However, as Filecoin’s network gradually emerged in front of people, problems gradually surfaced.
Filecoin governance dispute: towards centralization
The most motivating argument is that some of the actions taken by the Filecoin team during the testnet and the space race are full of a strong taste of centralization. Distributed storage blockchain is different from DApp platforms such as Ethereum. Filecoin may not have a strong developer ecosystem, but it cannot lack the support of miners scattered around the world.
However, in the face of the grand vision of building a “distributed network that stores the most important information of human society”, Filecoin chose to sacrifice the interests of some miners and even bear the reputation of “dictator”.
Effective storage and authenticated users
In Filecoin’s economic model design, the computing power obtained by storing the miner’s packaged data is divided into two types: original computing power and effective computing power. The so-called effective computing power refers to the computing power obtained by the miners storing the data of verified users (officially verified users), and the original computing power refers to the computing power obtained by storing data other than the verified users. Filecoin officially believes that the data of verified users is “valuable”, so it should be given more weight in the block incentive.
In the current economic parameter settings, the block weight for effective computing power is 10 times the original value. The transparent rules themselves are understandable. Filecoin officially plans to adopt the form of DAO organization to determine user qualifications in the future. However, this link is currently centralized.
Official hard drive Filecoin Discover
Filecoin released the Filecoin Discover project in May this year. Its original intention was to allow miners to store some useful data in the Filecoin network, such as human genes and meteorological data. Officials will store valuable and public data in the hard disk for sale. The data on these hard disks are marked as “valid data”, and the encapsulation of these data will obtain effective computing power. The block generation probability of effective computing power is 10 times the original value of computing power. In order to obtain more effective computing power, miners have to purchase the hard disk of Filecoin Discover in addition to encapsulating the data of verified users. So some people accused Filecoin of using its status as a “referee” for profit.
What is even more surprising is that the so-called “valuable” data is actually not valuable. According to the miner who received the Filecoin Discover hard drive, “The 8T hard drive claims that the data capacity is only 26G. The technicians can find a few files to read, and they can’t read any byte data.”
Miners suffered unfair treatment
On the eve of the space race, a miner with the pseudonym “Song Jiang” published an article called “Filecoin, A “Centralized” Storage Network” on the #fil-lobby channel of Filecoin slack, pointing out that the space race lacks fairness and There are problems such as the monopoly of miners and the people who issue cards.
During the space race, Filecoin miners complained endlessly. Due to the bugs and stability issues of the testnet itself, miners have to keep track of the official code dynamics at all times. If they don’t pay attention, they will be reduced their computing power and rewards.
Recently, the official announced a penalty measure for a few nodes in the competition. These nodes were officially recognized as cheating because they only packaged data for themselves. This incident was originally because the miners had drilled a loophole in the rules, and the officials transferred the consequences to the miners without communicating the new rules in advance. This arbitrary style ignores the interests of miners on the one hand, and exposes its lack of community governance spirit on the other.
Filecoin has been suffering for a long time. The miner community is quite dissatisfied with Filecoin’s strict economic model design and high mining threshold.
More important than the protests of the miners is that Filecoin has adopted a very complex system design in order to achieve effective storage. As a result, Filecoin can only develop with a single technical solution and economic model. The low scalability also limits the fault tolerance of the entire system. It can be said that once Filecoin has a big loophole in the technical or economic model, there is no possibility of turning back.
So, is there any good way to protect the interests of participants and promote the development of distributed storage networks? Pre-branch may be the antidote.
Undercurrents: bifurcation or inevitable
After pointing out a series of Filecoin problems, the miner “Song Jiang” said, “If the government continues to follow the current path and retains these centralized designs, it will be a serious violation of the spirit of the blockchain. If this is the final result, it must be Someone will fork the entire network. I will definitely be involved by then.”
Significantly, “Song Jiang” was originally the representative of resistance to centralized rule in “Water Margin.” This may be the voice of many miners who love and hate Filecoin.
However, forks are not a profitable business. It is inherently risky. Bitcoin is a lesson from the past. Success will bring fame and fortune, but if it fails, it will bear the reputation of splitting the community.
In fact, a better approach is to pre-fork. Start the fork chain before the Filecoin mainnet goes live. This will not only have no impact on the Filecoin network, but also provide references and references for the development of Filecoin.
We discussed the increasing centralization of the Filecoin team. In addition to avoiding centralization in governance, the forked chain should also improve Filecoin’s lack of mechanism design.
How to design a better distributed storage network?
To design a better distributed storage network, we must first understand the problems in Filecoin’s mechanism design.
Filecoin mechanism design flaws
Han Weiping pointed out the seven major problems facing Fliecoin development at the Xiamen World Congress:
- File hash value calculation is not friendly to Intel storage instruction set;
- The complex sealing process consumes too much computing resources;
- Low TPS limits the development of the number of network nodes;
- The pre-mortgage is too high, which is good for investors and not good for miners;
- The official data set restricts the use of the storage network, and the data security is not high;
- Frequent major technical changes;
- The blockchain economic model is unreasonable.
Han Weiping’s speech at the Xiamen World Mining Conference
In general, Filecoin mechanism problems are mainly reflected in two aspects: technical solutions and economic models.
Technical solution: In terms of storage proof, Filecoin uses complex copy proof and space-time proof to verify that the miner did store data and did not lose data during the sector life cycle. In data packaging, up to 11 layers of serial sealing calculations greatly increase sealing costs and time. The high cost of proof and sealed data not only wastes computing resources, but also raises the hardware threshold of the mining machine.
Economic model: excessive pursuit of effective storage, design of multiple mortgages, rewards and punishment mechanisms; excessive pre-mortgage is good for investors and not good for miners. And Filecoin officially has the authority to adjust economic parameters. This means that the government has become the hand of God in this economic system and can intervene in the operation of the economic system from a macro perspective.
In addition, Filecoin’s completely free data transmission method is contrary to the political and moral challenges of public information, and it is difficult to effectively supervise the content. This will lead to the rampant illegal content on the network and form an extra-legal place.
The first fork: Filecoincash inherits the ideals of IPFS
In fact, under the high demand for forks, some teams have begun to take practical actions to solve Filecoin’s problems. After analyzing the limitations of the current Filecoin network, Han Weiping announced the launch of the first forked project Filecoincash, using a different solution from Filecoin to implement the vision of the IPFS distributed storage network.
Filecoincash will improve Filecoin’s defects in seven aspects:
- The SHA256 algorithm is upgraded to SHA512;
- 11-layer serial sealing calculation is reduced to 5 layers;
- Add Layer 2 to improve TPS, support a wide range of applications such as smart contracts, and a market economy where the network scale can expand rapidly;
- Dynamic pre-mortgage to fully release the income to miners;
- Private data is determined by individuals to be effective and fully democratic, while public data needs to be reviewed and freely redundant;
- Joint development and governance with the community, fully democratically against dictatorship;
- Improve economic models.
Weiping Han said, “Currently, Filecoincash has conducted extensive communication with the top community participants and reached a consensus. It will open the Github code base on September 19, continue to iterate the code, open the test network synchronously, and launch the main network in October. “Invite the whole community to initiate and govern together.
Filecoin itself has a grand idea and design. It hopes to revolutionize the Internet giants’ control of user data through distributed storage technology, and is determined to expand the external utility of blockchain technology to “store important data for humans.” Both Filecoin and Filecoincash’s explorations are humanistic seeds planted in the soil of technology, allowing humans to retain privacy and enjoy justice while technology advances.
Finally, I share a scene described in the science fiction novel “Three-Body”:
Hundreds of years later, mankind is facing the crisis of extinction. In order to preserve the relics of earth civilization, mankind has built an earth civilization museum on Pluto to preserve earth civilization data for 100 million years. The question is, how to save the information? Scientists have discovered through various studies that the highest-end quantum memory of human science and technology has a shelf life of only 500 years, a USB flash drive can be stored for up to 5,000 years, and an optical disk can be stored for up to 100,000 years. These are too much for 100 million years. It’s trivial. The final conclusion is that the only feasible method known to store information for about 100 million years is to carve the words on the stone!
This clip reveals the importance of storage to human civilization. This idea may be far away, but who can tell?
Reference link
