History of Blockchain Expansion: State Channels, DPoS, Large Blocks and Sidechains

The original title: "The development bottleneck of Bitcoin and Ethereum is about to disappear. Taking stock of the expansion plans in the past 4 major directions, which ones are you optimistic about?" "

expansionexpansionIt has been 4 years since the term entered the public's attention.

In the past 4 years, countless teams have made extraordinary efforts for the expansion of the blockchain. Some of them are committed to improving the scalability of Bitcoin or Ethereum, while others simply start from scratch, wanting to achieve more without technical burden. Better and faster Bitcoin or Ethereum.

And withArbitrum's mainnet is online, and our book on "scaling", has turned to the penultimate chapter. Believe that after a few years withETH2.0 sharding + RollupIn the "final chapter", expansion will no longer be the main theme of the development of blockchain technology.

secondary title

01 Channel Lightning Network, Raiden, Celer, Connext

When it comes to expansion, channel may be the first word that pops up in the minds of many veteran players.

The earliest expansion technology should be regarded as the Bitcoin Lightning Network, and then there is a similar Raiden (Lightning Network) on ETH, including Celer and other projects later.

Lightning Network, as the earliest expansion solution, is undoubtedly a "failure" based on the development time and today's data. Raiden is the weakest one among the expansion solutions of ETH.

There is no other reason. The principle of channel technology is to open multiple micro-payment channels between the two parties of the transaction, connect each other and form a network. Most of the time, it is only suitable for high-frequency transaction scenarios with fixed transaction objects, and the channel nodes need to pledge the amount. It is very easy to form different routing bottlenecks and lead to transfer failure.

In 2017, there was a video explanation and derivation of Lightning Network that was very popular. With more and more channel nodes, those medium and large nodes, that is, nodes with larger pledge amounts, will become more and more popular because they do not It will cause a routing bottleneck. And those nodes with small amounts will be gradually eliminated.

It's very simple, if you want to transfer 10 BTC between A and Z through the lightning network, then the B, C, D and other lightning nodes on the way will be selected if the BTC balance is greater than 10, and those less than 10 will naturally be selected. open. If things go on like this, it is easy to deduce that the Lightning Network will slowly and spontaneously evolve and converge to the most efficient way, that is, several super-large nodes participate in almost all routes.

Um? How is this different from our traditional user-bank-user? Aren't these super nodes just banks? Is this still the main decentralized blockchain and Bitcoin?

Regarding channel technology, apart from Lightning Network, the two most noteworthy projects at present are Celer and Connext. The reason is that ETH’s current Layer 2 is severely fragmented. If it is impossible to form a dominant situation in a short time, then build a state channel between several Layer 2 The network, to a certain extent, alleviates this "fragmentation", but will become an indispensable technical solution.

secondary title

 02 DPoS EOS，Tron，BSC

EOS, Tron, as the most famous "Ethereum killer" in 2017-18, its technical core is undoubtedly DPoS, and many other high-speed chains of DPoS and DPoS variant series have been born later, and now the famous BSC is also DPoS used.

In fact, DPoS was first proposed by Bitshare in 2014, but it was EOS that entered the mainstream attention in 2017.

EOS has been controversial as soon as it was launched. Critics accused the design of 21 nodes as completely "not decentralized" in terms of quantity and hardware requirements. Those in favor of this one-coin-one-vote system The mechanism is most in line with the "representative democracy" in the political field of mankind at this stage, and retorted that the number of BTC PoW's current influential large mining pools is actually less than 21.

Objectively speaking, EOS has brought many new ideas to the blockchain, such as converting the CPU and memory of the entire network into resources that can be bought and sold. It is the first blockchain with a "constitution" and proposed an arbitration The concept of the court, the concept of "voting on the chain" is deeply rooted in the hearts of the people...

Of course, these have nothing to do with expansion, so I won’t show them for now.

In the field of capacity expansion, the biggest feature of DPoS is to allow 21 "high-performance super nodes" to handle all transactions, including packaging, block generation, verification and so on.

Here is a concept that needs to be emphasized, that is, "high-performance super node", which you will often see when we talk about other expansion solutions later.

DPoS is still PoS in essence, but it is different from ordinary PoS. In order to improve the performance and TPS of the entire network, DPoS has infinitely higher requirements for nodes than previous PoS, requiring nodes to have ultra-high computing performance and sufficient bandwidth. , to improve the processing capacity of the entire chain and the communication speed between nodes as much as possible, which is equivalent to eliminating all the short boards in the barrel theory in traditional PoW and PoS, and only using 21 boards that are basically equal in length and strong. Make this barrel.

The 21 nodes of DPoS are elected through voting on the Token holder chain.

Various variants of DPoS, consensus mechanisms such as LPoS, BPoS, and NPoS, as well as many new public chains behind - such as IOST, Algorand, etc., actually have the shadow of EOS, but the number of nodes is no longer 21. In DPoS With the removal of D, the way of generating nodes has changed from direct voting to delegated voting, or verifiable random function designation, etc. The requirements for node hardware are not as "perverted" as EOS, but high performance is still necessary.

It is very interesting that among all these "XPoS" at present, the most capable one is the youngest, with almost no originality, just fork ETH and then change the consensus mechanism from PoW to DPoSBSC, I don’t know if it’s a kind of irony.

secondary title

03 Big block BCH, BSV

Large blocks are a more "intuitive" expansion method than DPoS.

Since the TPS is so low now, the main reason is that there are too few transactions that can be accommodated in a block. Then I expand the block by 10 times, or even 100 times, from 1M to 10M, 100M, or even 1G. The TPS will not be corresponding Has it been enlarged by 10 times, 100 times, 1000 times?

The big block mainly revolves around the expansion of BTC and has nothing to do with ETH. It also gave birth to the famous blockchain in the history of the blockchain.BCH and BSV bifurcation eventimage description

(picture from the Internet)

Among them, BCH is regarded as the "moderate large block" genre, which believes that the block size should be gradually increased with the needs of the network and the technological progress of the Internet, while BSV is the "extremely large block" genre, which increases the block size from the beginning Later, the expansion plan of "unlimited blocks" was opened.

However, simply increasing the block size is not as simple as imagined. In the BCH stress test at the end of 2018, 16% of all nodes were offline and unable to synchronize data; in April 2019, the BSV stress test caused 6 blocks. reorganize. A series of problems such as node disconnection, block reorganization, and orphan blocks will gradually appear as the data of a single block (not the upper limit of the block) gradually increases.

Taking BSV as an example, in the concept of BSV, the final mature form should be a super node style similar to EOS, and even higher hardware and bandwidth requirements than EOS, almost at the data center level. In the end, through the survival of the fittest in the market, a few or more than a dozen "ultra-high-performance super nodes" will be left as miners. It's just not voting, but still maintaining pure PoW. The emphasis is not on the number of nodes or "decentralization", but on "Permissionless".

You see, high-performance nodes appear here again.

Whether the use of high-performance nodes will harm decentralization, the most important feature of the blockchain, is still a topic of debate, and large blocks are undoubtedly a front-row representative of high-performance nodes.

The two projects have also taken completely different paths in the direction of smart contracts. BCH first drafted the wormhole protocol, which was later declared a failure and went back to the pure electronic cash route. After BSC became popular, it began to develop a wormhole protocol called The side chain of SmartBCH, similar to BSC, is compatible with EVM. It is said that it has been basically completed, and I believe everyone will experience it soon.

secondary title

04 Sidechain Matic, Skale, BSC, Heco, Near, Avax, Fantom

Strictly speaking, the side chain itself is just a concept, not a specific expansion technology.

For example, the current Matic and Skale are side chains, both of which are PoS chains. Their consensus mechanisms are different, but they all serve ETH. Their initial purpose is to expand ETH.

The "trading platform chain" such as BSC and HECO directly forked ETH, and then the consensus was changed to DPoS. Many assets on ETH can also be crossed, so it can also be regarded as a side chain of ETH.

Near is sharding, Avax, Fantom is DAG... Their initial positioning is a native high-speed chain. But no matter what technical means, as long as they make a bridge to ETH, then it can be regarded as a side chain of ETH in a sense.

If a bridge is built on EOS and connected to ETH, then you can also say that EOS is a side chain of ETH.

Of course, the concept of sidechains is actually mutual. For example, Polkadot has a parachain like Moonbeam that is responsible for "bridging" ETH. At the beginning, there will definitely be many assets on ETH bridged to Polkadot. You feel that Polkadot can also be regarded as a sidechain of ETH.

But if one day the assets on Polkadot are richer than ETH, the market value is larger than ETH, and many native assets go from Polkadot to ETH through the bridge, can it also be said that ETH is Polkadot’s sidechain?

So the side chain itself is a mutual concept. It’s just that ETH is basically the only one at present, almost all valuable assets are in ETH, and ETH’s own performance is weak, so it makes people feel that all chains that “build a bridge to ETH” can be regarded as ETH’s side chains.

Each sidechain is itself an independent blockchain. Performance, security, etc. all depend entirely on the nodes of its own chain, and have nothing to do with ETH. Knowing the difference between side chains and Layer 2 is a key question to test whether you have a thorough understanding of blockchain expansion (here specifically refers to ETH expansion).

In the next part, we will continue to talk about several other mainstream expansion solutions, including Layer 2's Plasma and Rollup, sharding and parachains, and "non-mainstream" expansion solutions such as Solana, Flow, and Avax, so stay tuned.