In-depth dialogue: the status quo and future of fragmentation technology (2)

Decentralization, security, and scalability cannot be met at the same time, and it is called the impossible triangle of the blockchain. This conclusion is difficult to verify by means of mathematical proof, and has been questioned and challenged by many projects. Developers have proposed a variety of Layer 1 and Layer 2 expansion schemes for shards, DAGs, state channels, side chains/chains. The fragmentation technology improves scalability and does not reduce the degree of decentralization, and has a good application prospect.

Recently, several public chain projects such as Ontology have announced their own shard technology solutions. Some well-known shards such as QuarkChain have announced that they will be on the main online line in the near future, which has triggered the market's attention to fragmentation technology. TokenInsight invited QuarkChain, MultiVAC, Alephium, Ontology, Top Network and other public-chain projects using fragmentation technology to discuss the current status and future of fragmentation technology.

At the same time, "Dialogue Chief" also welcomes the chiefs of various blockchains such as exchanges, wallets, and mining pools, and participates in our activities to jointly contribute to the development of the industry. For more information, please contact TokenInsight Chief Communications Officer Vivi: tokeninsight_data .

The live event lasted for two and a half hours. The content shared by the guests was intensive and profound, so it was divided into two series to help readers.

The following is the [Dialogue Chief] text finishing version:

/ Third link: Round table 2 /

The chain is worth: My problem is that the fragmentation technology, especially the state fragmentation, is difficult to implement. So, is the fragmentation technology sufficiently strong for commercial availability? What problems need to be paid attention to during the actual landing of the fragmentation technology? Is there any possibility that future fragmentation technology will be replaced by other expansion technologies?

QuarkChain : If it is a large-scale commercial landing, then the fragmentation technology is a foundation in the blockchain . In terms of commercial availability, in addition to the fragmentation technology, there are other aspects, including Virtual machines, ledger models and token economics. Therefore, the fragmentation technology is considered to be a prerequisite and basis for commercial availability.

In the actual landing scenario, one of the core issues of fragmentation technology is the user experience. For example, from the perspective of the user experience, this technology is to make the user aware of the difference in the trading experience caused by the existence of the fragment, whether the transaction is an intra-chain transaction or a cross-chain transaction. It may not make a big difference to the user.

Then I think this is a key point of attention for fragmentation technology. It's like I use the Google search engine, the data he might use is distributed across different network nodes, but the user experience is a search result that returns a page.

The last question is that I personally think that it may not be replaced by other technologies, but it is combined with other technologies to better solve the problem of expansion. If there is a technology that can completely replace the fragmentation technology, then I think it may be a revolution for the entire distributed system. But at least in the short term it is difficult to see this possibility.

On the other hand, fragmentation technology and another DAG technology that we often refer to, these two are not conflicting technologies, and many of the fragmentation techniques are actually related to DAG, or the fragmentation technology is In a sense, it is also a DAG . So, if the fragmentation may be replaced by DAG, I think this may be a false proposition.

MultiVAC : I think the fragmentation technique is really difficult, but this is the best and most viable solution we have seen so far, and there are other better technologies that will not be ruled out in the future. But for the foreseeable future, we still can't see other better expansion plans for the time being.

There is less commercial availability. I don't think it should be eager to achieve success, but should think more clearly and do better. In fact, if you really want to use it, you can use it now, but the technology I have seen so far is not good enough, and it is not enough to meet everyone's expectations. For example, when the BTC paid, in fact, it already has a script, you can define what to say, I can unlock this UTXO money. However, when Ethereum was designed, its smart contract did not move to the BTC-compatible script. The experience of writing scripts before was actually useless.

Therefore, I think that in terms of commercial availability, the most difficult thing is to say no to the old things in the past. This is a very difficult and courageous thing. A good technical solution should lead the times.

Then I think that replacing Ethereum is not a better Ethereum. It should be a more advanced design. We think that sharding is the direction of the next generation blockchain. We should design a better development paradigm instead of thinking about it. Going forward compatibility , anyway, there is not much prosperity in DApp development, and then it is gambling in addition to ERC20.

So maybe this is the most difficult, that is to say no to the past. But goodbye to the old, we can usher in better, and then say no to the past technology. Nothing about the original knowledge of the industry. The more confidence you have in the future, the more patience you have now.

Alephium : Personally feel that fragmentation technology is very necessary, because once the user or usage comes up, the performance bottleneck will be visible immediately. For example, in 2017, we can experience it very intuitively in Ethereum. To the blockage of the network, once the blockchain is really used on a large scale, for example, a simple cross-border payment, the current single-chain system may not be fully supported.

Fragmentation technology, because it is more complex than single-chain technology, cross-sliced ​​transactions and cross-sliced ​​communication . These two technical difficulties will give the user experience a great impact on the system design. The project must consider these trade-offs when designing the application layer. For example, if a project uses two segments for its cross-segment transaction. The method of submission, then the design of his wallet and smart contract, will definitely deal with the two stages of submission. Some problems that may arise, such as delays, such as the successful submission of the first phase, and the failure of the second phase. How do you let the user know about this process at this time?

So here is the second issue, which is the user experience. On the one hand, we want to provide a black box so that users don't need to know our technical details. You don't need to know how the shards are done, you can use this technology; On the other hand, we have to provide some status related to our sharding technology, or some special data to the user, then how do users understand and how to use our application correctly. At this time, we need some time for the user to accept and understand the whole process of our technology. So I think this is actually a big difficulty. For example, mobile phones, from the initial Nokia mobile phones to Apple mobile phones, there are still a lot of people who are not very convenient to use.

Fragmentation technology coexists with other technologies. Because we first have no perfect consensus algorithm in the blockchain domain. In this case, different projects, different technologies will have different trade-offs, this time developers need to choose the most suitable platform for their own applications , for example, if an application requires higher transaction throughput, Then the speed requirement for cross-sliced ​​transactions is relatively high, so it is more appropriate to adopt our platform, because our platform cross-slice transactions are very fast.

Ontology : First, answer the question of whether the fragmentation technology has strong commercial availability. In our own R&D, there is a goal to better address the needs of to B applications. For the business of to B, each service in the blockchain may be implemented by means of a smart contract on the blockchain. Therefore, in the design of our fragmentation technology, it is basically a piecemeal design based on smart contract transactions. Another important issue in design is the security issue, because our goal is to do to B, so its security must be sufficient security.

In the actual landing, we need to pay attention to the problem, in addition to the technical aspects of the security and scalability of the fragment design, it is more important to choose a reasonable business model , so we chose this relatively simple way. It is the simple direction of to B that I think now. The other is about some commercial privacy protections, and our research is also relatively easy to implement in a fragmented network.

The last question is whether the fragmentation technology may be replaced by other technologies. I think it is possible. This kind of future is hard to say clearly, such as the security of cryptography, IP technology, etc. If it is suddenly obtained a comparison A big breakthrough, we think will have a big impact on the structure of the blockchain. But the sharding technology, I think and these technologies can be combined with each other , because these are not completely mutually exclusive foundations, so it should be a future direction for sharding.

Top Network : In fact, fragmentation technology is a rare case that can change business. After all , the main reason for blockchain business failure is throughput, real-time and certainty. If the fragmentation technology is properly applied, it can be solved. The bottleneck in the industry , there is really no mature solution before the filming, but this situation has changed. The head office of the industry already has a sharding solution that can be guaranteed to be safe, practical and can be landed. That is to say, what has been seen in technology can be commercialized.

From a technical point of view, the fragmentation and network fragmentation in the fragmentation technique does not involve state fragmentation and synchronization, and the implementation is indeed relatively easy. But the system that does not implement state sharding is not inherently capable of running. A very high TPS system, the block data of each day is massive. A system with a higher TPS, the data required to synchronize data per second is also extremely large, and eventually 99% of the nodes do not have such bandwidth. As a result of the evolution, either the system becomes a fully centralized, or semi-centralized system. That also violates a core value and philosophy of our blockchain.

Regarding whether the fragmentation technology will be replaced by other expansion technologies in the future, it should be said that there is no such thing as the replacement of the non-replacement. It should be said that there is a suitable and inappropriate distinction. However, from the perspective of computers, most of the methods and techniques for solving system throughput and performance in computer systems are actually generalized fragmentation techniques. From this perspective, fragmentation technology is the future solution to blockchains. The main means of industry performance and capacity.

/ Fourth link: Expert observer group question /

AnChain.ai : Many people think that the fragmentation technology solves the problem of insufficient expandability while ensuring the degree of decentralization, but it will affect the security. What do you think of the impact of fragmentation technology on security?

QuarkChain : First of all, I think there may be some misunderstandings about security. For the simplest example, the classic PoW chain like BTC, which defines security as a 51% attack. In reality, in addition to BTC mining using this algorithm, there are BCH, BSV, they are using the same algorithm for mining.

Because the BTC algorithm (SHA256) is diluted, the computational power required to attack him is now much lower than (all SHA256 algorithm power) 51%, not to mention BCH or BSV. Then our perspective on security issues is actually a question of attack costs . That is to say, consider the cost of a 51% attack just mentioned, such as electricity costs, rental costs and the personal impact costs of the attacker.

So, for example, let's assume that we have a fragmentation technique that allows digital certificates of the same algorithms, such as BTC and BCH, to exist in a new form of distribution. Everyone can share the power of BTC to ensure security. Then we think that they are actually more secure than the current way of adopting separate chains. So we think that here are some misunderstandings about the definition of security, especially the impact that fragmentation may have on security, not necessarily negative.

Alephium : Just now, Mr. Zhou talked about the safety issue during the expansion from the perspective of partial economics. I personally want to think about it from a more technical point of view. First of all, there should be a lot of people in China who have a saying that there is an impossible triangle. In fact, I think this should be a translation problem. It is actually a question of making a choice in three dimensions. Then when we are actually designing the sharding algorithm, we consider the dimensions of the trade-offs more. For example, we have to consider how much extra burden the sharding algorithm might bring to the system. Then consider, for example, what impact the contract design has on the fragmentation technology.

So different projects use different technologies, what kind of expansion can be extended to the end, and then how safe and decentralized, in fact, all dimensions are different. Then for our Alephium project, we value security very much. Our security is currently our algorithm guarantees that it is similar to BTC, basically the same, because we introduce DAG algorithm to slice The dependencies between the two are dealt with explicitly. In this case, if the user wants to attack a block or transaction on one of the shards, he must attack all other shards at the same time, which brings the security of 51% attack.

Dr. Wang Jiaping: What are the advantages of the sharding technique compared to other DELL (directed acyclic graph), sidechain and cross-chain, lightning network and other Layer 1 and Layer 2 enhancement scalability schemes?

TOP Network : This question, from a technical point of view, the main advantages of fragmentation are still three:

First, it can support relatively strong horizontal expansion and on-demand expansion, and different fragmentation technologies can be extended at different levels such as storage, communication, and computing. The blockchain industry is still in the stage where the performance bottleneck is far greater than the business bottleneck, so the support of the technology will be crucial.

Second, fragmentation technology is a technology that can be used in every level, without too many restrictions. We say that the main chain, side chain, cross chain, or shard can be used very well. In addition, since the core assets and state are on Layer 1, the value of the fragmentation technology is obviously higher than that of the off-chain technology.

Third, most of the means and techniques for solving system throughput and performance in computer systems are actually in the broad field of fragmentation technology. From this perspective, the blockchain industry can gain a lot of reference and inspiration from the traditional computer industry. The theoretical basis and practical experience of technology realization are also very numerous.

Of course, in different industries and applications, the focus can be greatly different. For example, some short-term high-frequency services are really the best match for the layer2 solution.

In addition, since I have talked about other expansion technologies, I have also introduced a few techniques:

One is that a purely chaotic structure of DAG does not lead to an extension of consensus. The difference between the DAG-Chain and the traditional Block-Chain core is that the former is to check the ticket after getting on the bus first, and the parallel writing to the chain is very fast, but the parallel verification of the complete validity and consistency of the transaction is not so fast; the traditional Block- The Chain is checked in after the ticket is checked in, and the validity and consistency of the transaction are packaged and checked in a fixed time interval. It is fair to say that DAG is called a million TPS. It is just a throughput indicator that is seen from the write and simple validity check. From the perspective of a transaction, the entire transaction process is finally confirmed (settled) for a short time, and the last transaction There is uncertainty in the timing of the confirmation.

The TOP uses the Lattice DAG (Lattice DAG). The Lattice-DAG is mathematically very orderly and consistent, which can guarantee the certainty and order of the transaction and avoid these problems.

Second, lightning network and state-channel technology do not perform well in most scenarios, and even cost more than the benefits they receive. Originally to speed up the transaction, but in fact the final locking and unlocking process is cumbersome, time cost is not small. Of course, we must also look at the application scenario. In the TOP Network, we optimized the State-Channel Layer 2 solution to solve the drawbacks of unlocking and locking, and also preserved the benefits of State-Channel in frequent transactions. It is especially suitable for decentralized distributed communication with high frequency and small amount.

The third is cross-chain and side-chain, which is actually a kind of fragmentation technology from the macro level. The middle layer of the TOP 3 layer book design is service-chain. It refers to the technology of side-chain.

Back to the original question, summed up: Fragmentation technology is the main means to solve the blockchain throughput and performance at this stage, but the fragmentation itself cannot solve all the problems, and must be targeted to the actual problems of the system. Other good technologies such as dot matrix DAG, layered network, Layer 2 and other technologies must be combined to innovate. Finally, do not limit your horizons. Generalized concurrency and parallel techniques and generalized network technologies are the next evolution directions of fragmentation technology. From the perspective of innovation, the main evolution of the blockchain should be toward comprehensive innovation, combined advantages, theory and engineering technology.

MultiVAC : I am a small peer with Jiaping . I used to be in an academic circle. I read Jiaping’s papers ten years ago. I don’t think everyone is doing the blockchain technology. This thing still needs to have the same technical background.

Let me first talk about my views on side chains, cross chains and multiple chains. In fact, this is a process of cognitive upgrading. Two years ago, when the film was still not fired at that time, many blockchain projects began to recognize that it was a light algorithm, including consensus algorithms, book structure, and tinkering. To improve performance, there is no way to really solve the problem of scalability. Therefore, everyone naturally thinks about the side chain and multi-chain schemes.

However, I think that one of the core problems of the side chain multi-chain scheme is that once multiple chains coexist between multiple chains, the number of miners in a single chain will definitely be insufficient, so the problem he brings is the safety of this chain. Sex is actually not guaranteed. This is the problem I just mentioned. This chain may always be such hundreds of miners or even hundreds of them. Just like now, BTC is only about 10,000 nodes, and then Ethereum is only about 14,000 nodes. If you want to build 100 chains, how many miners there are in the chain, and a few miners, its safety problems will be very serious.

In contrast, the strategy of “sequencing random distribution plus dynamic adjustment of miners” can effectively circumvent this problem. So in essence, everyone's thinking is the same, that is, using parallel solutions to form expansion. But the path to the solution is different, and everyone's consideration of security is different.

Then I will talk about DAG again. Today there are other projects that want to do DAG+ shards, etc. In fact, my key points are fierce, that is, I don't quite agree, DAG and shards can be superimposed.

For example, BTC and Ethereum are like a window in front of a window, the first one is in order, the queue is not very long, but very thick, everyone is squeezing to the window. Fragmentation is more like opening more windows, each of which can be processed in parallel. I think the difference between them is that the queues are all crowded forward. In fact, there is no fundamental problem of solving the processing power. In the DAG, there is a network storm problem and a book sorting problem. How do you solve it? I think these essential problems are difficult to solve, and the shards can really be processed in parallel. After the fragmentation technology really landed, I don't think I need to combine DAG.

Finally, I said the lightning network, and this problem including Layer2. The CTO of TOP NETWORK just said that all technologies are bordered and cannot cure all diseases. Then we know that the fragmentation can solve the problem of expansion, which is the Scalability problem. This can be solved. Teacher Jia Ping also said that in the fragmentation, the security of the fragmentation can also be solved. The problem that cannot be solved by the fragmentation is that if you want to guarantee decentralization, then the problem of transaction delay cannot be solved. . It is a transaction that requires a block of confirmation time. This problem cannot be solved by the decentralized solution of the slice, and the lightning network solves this solution very well.

We see the Lightning Network, which reduces transaction latency and boosts its throughput through the micropayment channel. This thing is very good, but there are some problems with the lightning network. Now look at the lightning network. If it continues to evolve forward, then it will definitely get the most efficient and optimal solution. It will become each of us, such as Alipay or a very centralized organization to build a micro. Channel, then through him to do this route. You will find that this solution is the last optimal solution for the lightning network, but this thing will become very central .

Including Layer 2, there are some risks of centralization and security, but he does bring some features that are not available in blockchain fragmentation technology. That is, it can upload additional information and value to the chain, which can achieve very low transaction delay, which is not available in the fragmentation technology. So I think the segmentation and lightning network includes Layer2. These things can be combined to some extent, and they don't conflict. But Layer 2 is less obvious for shards. Unlike Ethereum, its main chain performance is very poor, so Layer2 is extremely important. Sharding is relatively less important to the expansion of Layer 2.

This is my comparison and view of the above several programs. I don't think there is a need for DAG , sidechain networks and shards to coexist. But Lightning Network and Layer2 are still possible to further expand this segmentation . This is one of my views.

Dr. Jiaping Wang: I would like to summarize several aspects of performance: 1. Single-chain systems are not scalable, including algorand, and TPS is up to 3000, unless a third-party full-node is disabled as EOS becomes a cloud service; The cross-chain is scalable, but the one-chain-one service causes each service to be limited to the performance of the single-chain. It can only support the situation where each service is very small, but it is in line with the status quo, but cannot meet the big business in the future; 3, the side chain is the use of vertical business characteristics, for vertical business, but a Layer 1 can be deployed on a variety of Layer 2 sidechain, to make up for some of the missing Layer 1, such as transaction delay.

Dr. Pu Songtao: I have two questions: The first one is about the competitive situation of technological innovation. Fragmentation is an important frontier field of technological innovation in the current blockchain field. Blockchain researchers in various countries are actively exploring, you From the perspective of global technological innovation, what is the status of China in comparison with foreign countries in the technical segment of fragmentation, is it technology leading or backward? The other is about the ecological application of technology application. What preconditions do you think the application of fragmentation technology requires, and what preparations are needed from hardware equipment, system design, network conditions, etc?

Ontology : I think that although the country started late, it is at a similar level both at home and abroad. Every family has its own technical plan, and hundreds of people are vying for it. Unlike the previous years, the framework of Ethereum was basically adopted. However, the domestic aspect of innovation is indeed worse than that of foreign countries. This may be related to the domestic environment. The general prevalence of the project is relatively high. Usually, after the design is completed, it will be in the morning line, and overseas is more inclined to iterate.

In terms of the infrastructure required for fragmentation, if the blockchain is popularized in the future, the underlying layer should be implemented based on the cloud platform (AWS, etc.); on the landing, pushing the fragmentation to the B side is a reasonable direction. For Ontology sharding, since we are a multi-layer structure, each node is a Root Chain node and also a shard node. Therefore, the hardware requirements are relatively high. According to the number of nodes participating in the shard, the number of shards will be different. At least 4 core 32G hardware configuration is required, and there is definitely no problem in the cloud platform.

In addition, for the fragmentation, it brings about the improvement of TPS , the increase of transaction volume, and the increase of historical data . This is a problem that is particularly prominent in fragmentation. Storage separation should be an inevitable requirement and direction in the fragmentation network. There must be some storage node networks that provide storage for the entire network to prove such a service.

Finally, I think that after the large-scale application of the blockchain in the future, there will be a value-added service for the professional cloud service operation and maintenance of the blockchain application . Overall, I think the bottom layer is based on the current cloud platform, but I think there will be some Value-added services for blockchain applications.

View the previous article: