What is the relationship between IPFS and Filecoin? Why is IPFS delayed frequently?
The IPFS technology and the Filecion technology are actually two technologies. There are connections and differences between them. First of all, from a technical point of view, the technology stack used by IPFS mainly contains BT technology, plus other types of technology. It is an innovative project that integrates various mature technologies, which is IPFS. It should be noted that there is no blockchain-specific technology in the IPFS technology stack, such as consensus, such as tokens and so on. In other words, it does not use blockchain-related technologies.
Filecion is another technology, but Filecion and IPFS share the same module, which we call the basic module, called Libp2p module. The LibP2P module is both the network foundation of IPFS and the network foundation of Filecion. Filecion is a distributed storage project based entirely on blockchain. The technology it uses is blockchain technology. Filecion also has many innovative things, such as the proof of space-time POST, and it uses the data storage itself to perform computing power for consensus calculations.
The difference between Filecion and IPFS is the use of different technologies and different application scenarios. The scenario used by IPFS can be compared to HTTP, mainly for data transmission and positioning. What is transmission? Transmission is to transfer data from one node to another, that is to say from one place to another place. For example, when the mobile phone is opened on Jingdong, it is transmitted from the Jingdong server to your mobile phone. This is called data transmission, so one of the IPFS technical functions is to do data transmission. Another main function of IPFS is the positioning of data, that is to find where the data is, similar to HTTP DNS.
Filecion is mainly used for data storage. We know that data storage must be used for data transmission, otherwise your storage will be meaningless. So Filecion is actually an application built on IPFS. It builds a data storage project on top of IPFS, and their underlying networks can communicate. Therefore, data transmission and data storage are merged into a whole, and IPFS and Filecion form a set of complementary protocols.
IPFS went online in May 2015, and it has been five years now. Currently, Filecion’s second test is online. Most applications are now distributed in some private networks. There are not many public networks, and there are many top companies, including some large domestic companies, using IPFS technology.
IPFS is actually not much related to blockchain technology. It is just an Internet technology, but the underlying module used by IPFS is called LibP2P. After it is abstracted, it becomes an independent project, and this project is actually related to blockchain. There are quite a few connections, and it can be used as an infrastructure for blockchain projects. In other words, if a project wants to develop a blockchain project, it can be directly developed with the help of LibP2P, and a lot of work will be saved, which means that you don’t need to develop the basic network and data transmission parts on it. Because LibP2P has already packaged it for you, based on such a module, you only need to do the consensus part above to develop a blockchain project, which reduces the difficulty of blockchain developers.
IPFS is a protocol for distributed networks. The current network infrastructure is becoming more and more distributed. Whether it is a large network or a small network, the current microservices are actually distributed networks, including Jingdong, Taobao, Tmall, Alipay and WeChat are all distributed networks. The earliest centralized network was controlled by one node, and then many nodes, and then to a completely distributed network like the IPFS or blockchain era, so these are the two extremes of technology. IPFS is a technology born under the distributed network.
If you need to develop a distributed application, IPFS-based development is a very good choice, because the IPFS network itself has distributed characteristics. The IPFS node, also known as the peer-to-peer network, is a node that connects to any other node, which means that IPFS has a greater advantage over HTTP in data transmission. The core of IPFS is the BT protocol, so BT download is about 60% less bandwidth than the current HTTP for IPFS data download. This is the advantage of IPFS.
The specific application scenarios of IPFS can be divided into two types from a wide range: one is the public network application. The public network application is the large public network that we access the entire IPFS. It is like a huge BT network. We are exposed to the BT network. After that, you can develop some of your own applications. Most of these public network applications can develop some DAPP things or platform-related things; the other is private network applications. IPFS allows users to start an IPFS network, isolated from the main network. To form a small local area network, this kind of network that we can see more often used in enterprises. Now many companies are exploring this type of network in the enterprise intranet. As far as I know, some companies have started to use the IPFS network for their own projects.
Filecoin is a blockchain-based storage project. In layman’s terms, it is a huge network disk. This network disk is based on the blockchain project and is a network maintained and operated by miners. This is the popular understanding of Filecoin project. IPFS is the technical foundation of Filecoin, and some common modules are used between them. IPFS is a BT network and requires more nodes, because the more nodes, the better the performance.
In previous BT projects (for example, eMule), it is difficult to have a way to incentivize these people who use BT to open their own BT network and continue to seed others. Because it is not good for them, they also consume their own resources, such as power, hard disk and even network bandwidth. If you want to make the BT network better, you need more users to do seeds, but this is a very difficult thing. In the past BT network, it is difficult for us to let users actively share themselves The data.
We all know that the previous BT software has tried various incentive methods. For example, if you limit the bandwidth, then its software will actively limit your downloads. There are also some large BT stations many years ago. They use a point system, which means you can spend money to buy points. In fact, it is equivalent to spending money to buy traffic. This is a way to encourage everyone to do seeds. Not ideal enough.
But fortunately, we have encountered the blockchain in the past decade, and the blockchain has an inherent advantage in doing this, so Filecoin is an incentive layer above IPFS, and then can provide more to IPFS in the form of Filecion The nodes, this is the relationship between them. In other words, the more people IPFS uses, or the more applications it uses, the greater the demand for Filecion, the more Filecion nodes, and the greater the support for the IPFS network. They are a pair of mutually promoting protocols.
Although IPFS was launched as early as five years ago, in those days, because there was no incentive layer, when these companies or these developers used it, they could only store the data themselves, which was equivalent to fetching these IPFS nodes themselves. To store. After Filecion goes online, they have one more option, you can save the data without yourself, or even run a server node yourself, so it can continue the Filecion project, the two projects are combined to develop applications, this Is the connection between them.
Let’s talk about Filecoin’s consensus mechanism first. Let’s take the example of Bitcoin. It is well known that the consensus of Bitcoin is to use PoW, which is to use the computing power in your hands as the proportion of votes. For example, if you have 1% of the computing power of the entire network, then in terms of probability, you will have 1% of the Bitcoin return of the entire network. This is the consensus mechanism of Bitcoin, that is, the mining model. The computing power it uses is actually PoW proof of work.
In fact, Filecoin can also use bitcoin. The white paper has made it very clear that Filecoin can use any type of consensus mechanism. As for which consensus mechanism you choose, it is entirely a team’s choice, and there are no mandatory rules. Why Filecoin does not use the same consensus mechanism as Bitcoin. We all know that Filecoin is a project based on the real economy. It is very closely integrated with the real economy, so it must be reduced in cost. Just said that it is a huge network disk. If it does not reduce costs, it means that the network cost is too high. I believe no more people will choose.
Based on this understanding, Filecoin must find a consensus mechanism, the cheaper the better, this consensus mechanism obviously cannot use Bitcoin’s consensus mechanism, because Bitcoin’s consensus mechanism is very expensive, and Filecoin needs to circumvent this problem. We just said that Filecoin is a huge network disk, and what the miners do is to provide their own storage and bandwidth. Can we change our thinking? I take the storage space you provided, that is, how does the data you store itself be used as the basis of your consensus mechanism? This is the direction of Filecoin.
Based on this thinking, the Filecoin consensus mechanism proposes a spatio-temporal proof called PoST. The spatio-temporal proof is to prove that a miner stores a certain amount of data for a period of time. As long as you store the data, you have computing power. It is essentially the same as Bitcoin. For example, if the size of the data you store accounts for 1% of the total network data, then the Filecoin return you get, that is, the token return is also 1% of the entire network. So it is based on the amount of data stored. Once we calculate this consensus mechanism according to the amount of stored data, obviously no huge calculation work is required to complete this thing, we only need appropriate calculation work to complete, which can save computing resources and power.
This is Filecion’s efforts to reduce costs. For Filecoin, it will face huge challenges. No one has done this before, this innovation is very challenging, which means that Filecion’s consensus mechanism is too difficult, because it has to stare at the data itself stored by the miner, the miner cannot delete the data, it is not possible to temporarily generate data .
The data itself is in Filecion, because we have just said that Filecion is a huge network disk, and this network disk is used to store data. The data comes from the user, from the user. The user gives his data to the miner, and then the miner takes this data and stores it as their own computing power, and then maintains the consensus mechanism of the blockchain to obtain the reward of the blockchain. This is the consensus mechanism of Filecion mining. What is particularly clever here is that Filecion can use the data stored by miners when it does its consensus mechanism.
So the consensus mechanism in Filecion, those mentioned above, this is just a metaphor, here only involves the concept of a proof of time and space, there are many other concepts that are not involved, just to let everyone understand what it is doing What kind of thing can be.
There are probably several roles in Filecion’s ecosystem. One is the developer, because Filecion is a blockchain that can be developed on, and it is a blockchain combined with the real economy. In this way, it means that the developer is to do the application on it, otherwise Filecion will lose its own value, so the developer is one of the roles here. The other is a miner. In the Filecion project, there are two types of miners. One is called the retrieval miner and the other is the storage miner. I will use a simple description to tell them what they are. It should be clear at a glance.
In the traditional BT project, when you download data, do you particularly want someone to help you seed? In this way, you can download the data. I believe some people will encounter a situation like this, that is, when you download a file and download 50%, it turns out that there is no seed and your download is interrupted. In Filecion. The concept of a miner is to make a seed node for everyone, and then when you download, you can download from many places, there will be network acceleration, this is the miner.
We just mentioned developers and miners. There are also coin holders. It is very similar to Bitcoin holders, investors, enthusiasts or exchanges. But Filecion’s token is not quite the same as Bitcoin’s token. Filecion’s token has transaction value. Because in this network, we have to measure how much money is stored or how much money is downloaded, this thing is measured using tokens.
Everyone must pay attention that every bit on the Internet is paid for, and there is no free lunch in the world. For example, if you open Jingdong on your mobile phone, and then you download the product data of Jingdong, then I want to tell you what the payment process is. First, you need to pay the traffic fee from the network operator, and then download the traffic to your mobile phone , This traffic is paid. Secondly, then these are stored on JD’s servers, and JD.com has to pay storage fees. Finally, when downloading data from JD.com, JD.com will pay you another traffic fee. In fact, this traffic is in our current economic model. It is charged in both directions. So all data is paid by someone, and all storage is paid by someone. The economic model of Filecion fits exactly with these laws. A user is a role, it is a user of Filecion’s network.
Finally, let’s talk about the concept of the mining machine in Filecion. The concept of the mining machine in the Filecion system is not the same as Bitcoin. The mining machine’s computing power has been determined when the Bitcoin mining machine leaves the factory. We are used to the concept of computing power Describe the mining ability of a mining machine, this is the computing power. When the Filecion mining machine leaves the factory, its computing power is zero. You can’t predict how much computing power it has. then what should we do? What is its computing power? Its computing power is calculated in this way. How much effective data do you store in it? The computing power is measured by the amount of data you store. For example: the data storage capacity of the entire network is 100T, and if you save 1T yourself, then the computing power of your mining machine is 1%. If after a while, the data volume of the entire network becomes 200T, and your data volume has not increased, then you have changed by 0.5%. Therefore, its computing power is dynamically changing, which is not the same as Bitcoin. You should not measure the computing power of Filecion miners in this way.
At present, the Filecoin test has been going on for a small period of half a year. Due to the impact of the epidemic, an appropriate test extension period has also been made. In fact, there have been many problems in the entire testing process. For example, the network reset, the test is to find out the bugs of this test network.
The test has been divided into three stages so far. The first stage is the development network test. In fact, it is the development and testing. It is an open source project and requires a large development network. The test has no reference value at all. The first stage test is During the test period of two or three months before today, the development team improved many things and changed some agreements. Because when testing, I gradually found that there are some problems that need to be compromised and weighed, and a game between security and cost is played. There are many changes in terms of technology and economic models.
There are still some changes between the first phase of the test and the launch of this test today. This is something that cannot be presented or guaranteed in the previous economic model, but we found that this problem must still be solved. In theory, the launch time of Filecion should be between July and August, and there will not be too many problems, because the second stage of the test, the network has been much more stable than before, including the use of here Some new practices. There are two implementations of Filecion, one is known as Lotus, and the other is go-filecoin. Now that the network has achieved interoperability, the network has been connected, and the second stage will be two of the two networks. Comparison of implementation methods may result in two kinds of coexistence, or only one kind will be used in the future.
Filecion has progressed to today, and many people are not sure about its difficulty. I always think that the Filecion project has been dragged on and off, but in fact it is because the difficulty of this project is too great. It involves a lot of frontier knowledge of mathematics. For example, zero-knowledge proof or non-interactive zero-knowledge proof. Since this knowledge is not perfectly solved in mathematics, Filecion is a semi-scientific project to a certain extent. There are many things that are doing scientific research, so its project plan has not been completely stable. Now many universities are participating in the research of this project. During the second test, we can see some currently used proof mechanisms or some mathematical mechanisms.
Its stability is already quite good, but it is not ruled out that there may be changes, including the economic structure and technical structure of the main network. Let’s put it this way, about 80% can go online on time in July and August, but there is a very small probability that it will be delayed later, such as a very serious bug. Therefore, it is appealed to miners or mining machine manufacturers to participate in the test and leave the bug before it goes online. It must not stay after it goes online. That may be catastrophic for everyone.
Participating in Filecoin mining is not the same as Bitcoin. Let’s use Bitcoin as an example for comparison to make it easier for everyone to understand. We all know that Bitcoin’s earnings curve is particularly easy to calculate, because what is the cost of electricity? What is the management cost? Including machine room costs, mining machine costs. So it can accurately calculate the rate of return. Bitcoin mining is particularly transparent, and it is easy to figure out how much investment can be made, including risks, so this is Bitcoin.
But Filecoin is not the same, it is relatively more complicated, and it needs some technical support for mining. If it is bitcoin, buy a mining machine and plug it in, you can mine. Filecoin is different. If you buy a mining machine and plug it in, it may not be able to dig up a bit of currency, because we all know that it has been said that its computing power is stored data. You must pay attention to you How much data is saved. So it still has a certain technical content and technical threshold.
To participate in mining, the first is to learn enough knowledge. Miners must understand how this system works. When you understand how it works, you can make a profit curve that you estimate in the future. , So you have to prepare sufficient knowledge.
Second, if you feel that learning this thing is too difficult, there is another way. I suggest that you can find a team with stronger technical strength to invest. Of course, you must pay the toll. In the end, it depends on the rate of return, the most common investment project.
Therefore, if you participate in mining, it is very important not to invest prematurely. This is something I have been reminding everyone since 2017. Now that it has reached the second test, I still tell everyone not to do prematurely. Invest, you can observe again. Let’s take another look at the actual measurement situation. Because Filecoin’s current network has not reached the final stage of code stability, it is not clear what the final form of a series of things like Filecoin’s data mining performance will look like.