In the blink of an eye, four years have passed since the word “expansion” entered the public’s attention in 17 years.
In the past 4 years, countless teams have made great efforts for blockchain expansion, some of them are dedicated to improving the scalability of bitcoin or ethereum, while some of them simply start a new kitchen, trying to achieve better and faster bitcoin or ethereum without technical baggage.
With the launch of Arbitrum’s main website, we have reached the penultimate chapter of the “Scaling” book. I believe that in a few years, along with the “final chapter” of ETH 2.0 Slice + Rollup, scaling will no longer be the main theme of blockchain technology development.
Before the final chapter, let’s briefly review those scaling solutions we have chased together in the past few years.
Lightning Network, Raiden, Celer, Connext
When it comes to scaling, channel may be the first word that pops into the heads of many veteran players.
The earliest scaling technology should be considered the Bitcoin Lightning Network, and then there was the similar Raiden (Thunderbolt Network) on top of ETH, including later projects such as Celer.
Lightning network as the earliest scaling solution, taking the development time and today’s data, is undoubtedly considered a “failure”, Raiden in the ETH scaling solution is the weakest presence in one.
The reason for this is that the principle of the channel technology is to open multiple micro-payment channels between the two sides of the transaction to connect with each other and form a network, which is most of the time only suitable for high-frequency transaction scenarios with fixed objects, and the channel nodes need to pledge the amount, due to the different pledge amounts, it is very easy to form different routing bottlenecks and lead to transfer failure.
The lightning network has a video explanation and derivation in 17 years is particularly hot, as 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 cause routing bottlenecks. And those nodes with small amounts will be phased out.
Very simple, if you want to transfer 10 BTC between A to Z through the lightning network, then the route of the B,C,D and other lightning nodes, whose BTC balance is greater than 10 will be selected, less than 10 natural to bypass. In the long run, it is easy to deduce that the lightning network will slowly and spontaneously evolve to converge to a most efficient way, that is, a few oversized nodes involved in almost all the routes.
Hmm? How is this different from our traditional user-bank-user? Aren’t these mega-nodes the banks? Is this still the main decentralized blockchain and bitcoin?
The reason is that ETH is currently severely fragmented by Layer2, and if it is not possible to form a dominant situation in a short period of time, then building a state channel network between several Layer2s will alleviate this “fragmentation” to a certain extent, but will become an essential technical solution.
But in any case, the channel technology is destined to be used only as a capacity expansion program “auxiliary position”, not to stand in the front row, let alone the C position.
EOS, Tron, BSC
EOS, Tron, as the most prestigious “ethereum killer” in 17-18 years, the core of its technology is undoubtedly DPoS, which has spawned many other DPoS and DPoS variant series of high-speed chains, and now the famous BSC, which also uses DPoS.
DPoS was actually first proposed by Bitshare in 2014, but it was EOS that entered the mainstream in 17 years.
EOS has been controversial since its launch, with critics accusing the 21-node design of being “decentralized” both in terms of number and hardware requirements, while those in favor of it believe that this one-coin-one-vote mechanism is most in line with the current stage of “representative democracy” in human politics, and countering that the number of influential large mining pools in BTC PoW is actually less than 21.
Objectively speaking, EOS still brings a lot of new ideas to the blockchain, such as converting the CPU and memory of the whole network into resources that can be bought and sold, being the first blockchain with a “constitution”, proposing the concept of arbitration committee court, and making the concept of “voting on the chain” deeply popular ……
Of course, these have nothing to do with scaling, so let’s put them aside for now.
The biggest feature of DPoS in the field of scaling is that it allows 21 “high-performance super nodes” to handle all transactions, including packing, blocking, validation, etc. Here is a concept to emphasize in particular, namely, the concept of the “high performance super nodes”.
Here is a concept to emphasize, that is, “high-performance super nodes”, later when we talk about other scaling solutions, you will often see.
DPoS is still essentially a PoS, but different from the ordinary PoS, in order to improve the performance and TPS of the whole network, DPoS requirements for nodes than the previous PoS countless times higher, requiring nodes have super high computing performance and sufficient bandwidth, as far as possible to improve the processing power of the whole chain and the communication speed between nodes, equivalent to the traditional PoW, PoS inside the barrel theory of all the short boards are eliminated, only 21 basically equal length and strong boards to do the barrel.
The 21 nodes of DPoS are elected by Token holders through on-chain voting.
The later variants of DPoS, LPoS, BPoS, NPoS and other consensus mechanisms, as well as many new public chains – such as IOST, Algorand and so on – actually have shades of EOS, except that the number of nodes is no longer 21, the D in DPoS is removed, and the nodes are generated by proxy voting instead of direct voting, or by verifiable random function designation, etc. The hardware requirements for nodes are not as “sick” as EOS, but high performance is still a must.
It is interesting to note that the most capable of all these “XPoS” is the youngest, almost nothing original, but simply Forked ETH and then changed the consensus mechanism from PoW to DPOS BSC, I wonder if it is an irony.
But anyway, high-performance nodes from the DPoS era of EOS, is officially into the blockchain “miner” class, not only tends to centralize, but all blockchain projects, there will be.
Big Block is a more “intuitive” way to expand capacity than DPoS.
Since the TPS is now so low, the main reason is because a block can accommodate too few transactions, then I expand the block 10 times, or even 100 times, from 1M to 10M, 100M, or even 1G, the TPS will not be expanded 10 times, 100 times, or 1000 times accordingly.
The big block is mainly around the BTC expansion, not related to ETH, but also the birth of the famous BCH and BSV fork in the history of the blockchain, the three camps war of words for more than a few years, basically who can not convince who, quite a sense of the “Christian” fork.
The BCH is a “moderate big block” school of thought, believing that block size should be gradually increased in line with the needs of the network and the technical progress of the Internet, while the BSV is an “extreme big block” school of thought, which has raised the block size to very large from the beginning, and later opened up the “uncapped block” expansion program.
However, simply increasing the block size is not as simple as one might think. In the stress test of BCH in late ’18, 16% of all nodes were offline and could not synchronize data; in April ’19, the stress test of BSV resulted in the reorganization of 6 blocks. A series of problems such as node dropouts, block reorganization, and orphaned blocks will gradually become apparent as individual block data (not the block cap) gradually increases.
Take BSV as an example, in BSV’s vision, the final mature form should be a super node style like EOS, with even higher hardware and bandwidth requirements than EOS, almost at the data center level. Eventually, through the market’s elimination of the fittest, a few or a dozen “ultra-high performance super nodes” will be left to be miners. Only instead of voting, it remains pure PoW, emphasizing not the number of nodes or “decentralization”, but “Permissionless”.
You see, high performance nodes, here they are again.
Whether the use of high-performance nodes hurts decentralization, the most important feature of blockchain, is still a topic of debate, and Big Block is certainly a front-row representative of high-performance nodes.
The two projects also took completely different paths in the direction of smart contracts. BCH first started drafting the wormhole protocol, then announced its failure and went back to the pure electronic cash route, and then after the BSC fire, started developing a side chain called SmartBCH, similar to BSC, taking the EVM-compatible route, which is said to be basically complete, and I believe we will soon be able to experience it.
The company’s competitors are the various sidechains of ETH and L2, while BSV has taken ETH itself as a competitor from the very beginning. It is believed that there will be at least three to five years before the market will give a relatively definite answer as to whether this expansion method of large blocks is feasible or falsified.
Matic, Skale, BSC, Heco, Near, Avax, Fantom
Strictly speaking, sidechain itself is just a concept, not a specific means of scaling technology.
For example, the current Matic and Skale are side chains, both are PoS chains, their consensus mechanism is different, but they both serve for ETH, and their initial purpose is to expand ETH.
BSC and HECO are “trading platform chains” that directly fork ETH and then change the consensus to DPoS, so many assets on ETH can be crossed over, so they can also be considered as side chains of ETH.
Near is a slice, Avax, Fantom is a DAG …… They are positioned as a native high-speed chain at the beginning. But regardless of the technical means, as long as they make a bridge to ETH, then they can be considered as ETH’s sidechain in a sense.
If a bridge is made on top of EOS to connect to ETH, then you can also say that EOS is a sidechain of ETH.
Of course, the concept of sidechains is actually mutual. For example, there are parallel chains like Moonbeam on top to “bridge” ETH, so at first there will be many assets on top of ETH that will be bridged to PoC, and you feel that PoC can be considered a sidechain of ETH.
But if one day the assets on the boca are richer than ETH and the market value is bigger than ETH, and many native assets go from the boca to ETH through the bridge, can we also say that ETH is the sidechain of the boca?
So the sidechain itself is a mutual concept. At present, ETH is basically the only one, almost all valuable assets are in ETH, and the performance of ETH itself is weak, so it makes people feel that all the chains that “take the bridge to ETH” can be considered as side chains of ETH.
Each sidechain itself is an independent blockchain. Performance, security, etc. all depend on the nodes of their own chain and have nothing to do with ETH. Being clear about the difference between sidechain and Layer2 is a key question to test whether you have a thorough understanding of blockchain scaling (specifically ETH scaling here).
In the next article, we will continue to talk about several other mainstream scaling solutions, including Layer2’s Plasma and Rollup, sharding and parallel chains, as well as Solana, Flow, Avax and other “non-mainstream” scaling solutions, so stay tuned.
Author | The Five Fireballs
Posted by:CoinYuppie，Reprinted with attribution to:https://coinyuppie.com/bitcoin-ethereums-development-bottleneck-is-about-to-disappear-inventory-of-the-past-4-major-directions-of-expansion-solutions-which-ones-are-you-optimistic-about/
Coinyuppie is an open information publishing platform, all information provided is not related to the views and positions of coinyuppie, and does not constitute any investment and financial advice. Users are expected to carefully screen and prevent risks.