The PoS algorithm does not prove to be more secure than the PoW

The original title:Nervos: the future of blockchain systems must be layered

Original author: babbitt Helen

On May 16, directed by hangzhou financial services office, babbitt week block chain, the host of the 2019 hangzhou Chainge technology on the opening day, Nervos COO Daniel lv to the Nervos CKB chain blocks the underlying infrastructure next generation "as the theme brought to share, he said: only two direction and future development of the chain, one is called chain expansion, expansion and a call outside the chain. Expansion on the chain is to make a stronger chain, but inevitably in a chain to make trade-offs and trade-offs. According to lv guoning, the core value of blockchain is to provide trust. In the future, the blockchain system must be layered to solve different problems at different levels, so as to take the road of gaining trust on the chain, achieve higher performance outside the chain, and finally achieve the maximum effect in a holistic way.

Here are edited excerpts from the live speech, compiled by babbitt.

Today is technology development day, I will share three points, Nervos CKB expansion plan, design methodology and technical features.

& have spent 1. Capacity expansion on or off the chain?

Have you noticed that at the end of 2017 and in 2018, there was a wave of public security chains. These public security chains have all kinds of technical directions. Not only you, but also I don't know the technology of these chains. I don't understand them.

But if we look at the convergence from a higher point of view, there are only two directions for the future development of the common chain, one is called expansion on the chain, and the other is called expansion on the chain.

1. Capacity expansion on the chain

There are several ways to expand the chain. The first way is called "bigger block", which is what the BCH is doing. The second way to expand capacity on the chain is to "use faster consensus algorithm", because it is found that the performance of a chain is not enough and the TPS is too low, so the increase of consensus nodes will lead to the bottleneck of performance processing.

We are in line with the principle, when we see a chain where there is a performance bottleneck, we use a more powerful technology to put it in. For example, some people think that block chain is a single line system, so there is the third way "parallel computing-sharding technology"; Some people think that the fourth way "DAG", a new topological data structure, is more suitable for processing massive data and massive transactions than the chain structure.

In fact, all the chain expansion plans are doing one thing, is to make a stronger chain, let this chain solve all the problems. But if you are a researcher, you will find that there are some problems behind the expansion scheme on the chain.

For example, "larger block", when your TPS performance into a hundredfold, thousands of times magnification, will immediately encounter a new problem, storage space explosion; For example, "use faster consensus algorithms." now we have POS and DPOS. Do they compromise on security? Let's look at whether or not these algorithms compromise security, not to mention the decentralization that these algorithms create;

For example, "parallel computing-sharding technology", parallel algorithms add several dimensions to the complexity of the implementation of the chain, we ask a question, such a complex function implementation, how do you ensure that the whole system is absolutely safe and reliable? After all, it hasn't stood the long test of reality, and complexity is the enemy of security.

For example, "DAG" is more from the academia, but in fact, DAG technology has not been really recognized by the academia, it is now landing and processing a large number of transactions, there are still many problems rely on the centralized approach and solutions.

So in retrospect, the way we expand the capacity of a stronger chain is essentially to find ways to make all kinds of trade-offs within a chain, to gain performance, but what do you lose in the process? This is a problem.

2. Chain expansion

The other way is called chain expansion. We see that there are many out-of-chain expansion schemes, such as lightning network, state channel, Plasma, Truebit, etc. However, there is also a problem with out-of-chain expansion:

For example, if you buy a mobile phone today, the person selling the mobile phone tells you that all the mobile phone applications released before today can be installed and used, but the APP released after today may not be supported by your mobile phone, do you still want to buy this mobile phone?

Block chain chain expansion has such a problem, now all the two-layer expansion plans, they can cut user experience, cut TPS, cut high performance, but the biggest bottleneck comes from a layer of network does not support them.

How do you get a layer of network to support them? It's technically possible, but you have to do a million hard forks to support it step by step, which is a tricky process because blockchain is not a piece of software and is very hard to upgrade. Every time I do a hard fork, I have to do a lot of work, and I can only move forward in small steps. Now with all tier 1 networks, bitcoin, ethereum, etc., it's very difficult to support tier 2 networks.

& have spent Ii. Block chain design methodology & NBSP;

After the above thinking, we have a set of block chain design methodology, mainly including the following four things.

1. What is the core value of blockchain?

What should we do in the future? The first part of the chain expansion or chain expansion? We think the core of the question goes back to the essence: what do we really want to do with blockchain? Or what is the core value of blockchain?

The core value of blockchain is to provide trust. That's the conclusion. I won't break it down. Based on this, we can draw a corollary: blockchain helps you achieve trust through global consensus. And the global consensus, it's a slow, expensive thing, you have to accept it, that's the truth.

The consensus of distributed, decentralized systems cannot be done faster than centralized systems. The core of a distributed system is to give you trust, and trust is its core value. When we do blockchain, we should consider how to bring out the value of trust.

2. The block chain system will be layered in the future

If you had a system out of the chain that could improve your performance on a safe basis, would that be possible? You can, of course, as long as you design LAYER1 in the direction of security, and then pass the trust outside the chain to secure the system outside the chain.

When we think about this, we come to the conclusion that instead of making trade-offs on one chain, we go down the path of gaining trust on the chain and achieving higher performance outside the chain. After a period of verification, we believe that the future of block chain is to layer, divide the network into different levels, solve different problems by different levels, and finally put these levels of system together to achieve the maximum effect in a holistic way.

Consider different levels of needs

If the blockchain of the future is a hierarchical network, then we need to think about the roles of each level in that network.

Needless to say, its core role is to provide performance, to provide user experience, to cut scenes, to solve businesses, and it does that very well, it just needs LAYER1 support.

Of course LAYER1 should not only pursue performance, in the block chain scenario, LAYER1 as long as the ultimate security and the ultimate decentralization to ensure the security of the upper application, the system will run.

4. The design of the underlying protocol is fully optimized for the upper layer

With that in mind, when we figure out the relationship between LAYER1, LAYER2, and the characters, we have to rethink what we're doing when we're doing the underlying design.

The bottom thing to do is to fully optimize for the top without sacrificing security and without sacrificing decentralization.

So here's a graph, very simple, with a line in the middle, LAYER2 on the line, LAYER1 on the line, that's going to represent a performance with a circle. There is no problem with the expansion of the line. Under the dividing line in the middle, as long as we focus on DECENTRALIZATION and SECURITY, it is DECENTRALIZATION and SECURITY.

& have spent Three, Nervos CKB technical characteristics & cake;

Based on such thinking, we think that LAYER1's bottom layer is to do the following things:

1. Asset management and storage model

First of all, between LAYER1 and LAYER2, you want to ensure that security and security can be passed up, and its core thing is to ensure that the assets are secure. To ensure the security of assets, it is not only necessary to ensure the security of platform Token, but also to enable users to issue custom assets on the platform and protect the security of user-defined assets.

You can ask a question, we send an ERC20 Token on ethereum, maybe this Token becomes valuable for some reason, who guarantees its security? That's interesting. We'll get to that in a minute.

Second, it is necessary to support the user through a custom function, and in the underlying model design books, should be based on assets as the guidance, because only the assets oriented to design book model, it is more suitable for can be mapped to the assets of the state of the user of the upper application do trading, do business, network, finally also are more likely to return to the underlying make liquidation.

So we borrowed the idea of UTXO of bitcoin and designed our own independent and very characteristic ledger model system, which is not expanded here.

In addition, in addition to issuing assets and insinuating assets, when LAYER2 users need to return assets to the bottom for network liquidation, they will submit cryptographic evidence, and you need to have storage capacity for the data of cryptographic evidence, so LAYER1 must do a good job in real estate management and general storage model.

2. The CKB - VM virtual machine

The second is virtual machines, which are required by all blockchain businesses. You want to figure out what is the positioning of the chain, is it a high-level language support virtual machine in the application layer, or is it a security system at the bottom of the virtual machine?

We think at the bottom of LAYER1, the core is to make the virtual machine secure enough, flexible enough, and have a lot of runtime transparency, so you need to make sure that every instruction that you execute in the virtual machine is very easy to calculate its cost, which is very important for your underlying economic model. So we used a lower-level abstraction, called the risc-v hardware CPU's instruction level, to simulate the implementation of our virtual machine, which was executing like a hardware CPO.

3. Nc-max PoW consensus algorithm

The underlying security, we use the PoW consensus algorithm. Now many people who discuss PoW and PoS algorithm more secure, and no way to prove the truth PoS algorithm is more secure than the PoW, but it significantly more complex than the PoW, the complexity, there is no way to guarantee the design and engineering realization of the theory can do safety, even if it project implementation under the condition of relatively complete, also has not been in a production environment for ten years.

So we thought the PoW algorithm was the better choice, but we didn't stop there. We're going to squeeze every last drop of hardware out of the way to make the performance fast enough, and that's what we're doing. We have proposed a consensus algorithm of nc-max, which will be sent out next month. We will have a very clever way to squeeze the bandwidth utilization to improve the efficiency of the algorithm, while the algorithm of PoW is how to efficiently use the bandwidth instead of wasting it.

4. Economic models

Finally, and most importantly, the economic model. When your network is divided into LAYER1 and LAYER2, you have to think about how you can capture the value generated from the previous layer to the bottom as your upper application becomes very prosperous, and when the bottom value is captured and deposited, it can motivate your miners to get more benefits. When miners get more revenue, it is also the algorithm that allows miners to get more resources to ensure the whole network. When the security boundary of the whole network is improved, it can also provide better security protection for the upper layer, which is the core of the economic model design.

The core of economic model design is not to encourage every consensus. The core of economic model design is to promote the overall ecology to be more and more prosperous.

Finally, let's share what we are doing in 2019. We will release the economic model design proposal on March 8, release the test chain on May 18, and release the consensus algorithm Paper in June, which is expected to launch the Q4 main network.

Click to view the original article

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia