Written on the eve of Ethereum's "The Merge"

In the morning, I watched V God’s live broadcast on ETH Shanghai, and I would like to share some new insights with you.

1/
First of all, what we can see is that V God deliberately downplays the concept of "ETH 2.0", and puts more emphasis on specific technologies such as PoS and Danksharding.

2/

In fact, it is not only Vitalik who can perceive that the term "ETH 2.0" is deliberately avoided by the entire Ethereum community to prevent misunderstandings by others. To put it bluntly, the "merge" of PoS ETH and PoW ETH is a simple consensus mechanism switch, more like a power transfer, and there is no substantial change at the user level.

3/
After the "big merger", all PoW mining machine nodes were offline and replaced with PoS nodes, such as Lido, CEX and so on.

To be honest, the difficulty bomb is here, and graphics card miners can't afford it unless they engage in a fork. However, the current speaking power and time window of ETH PoW miners are not enough to support such a fork.

4/
Therefore, PoS ETH is no longer a new chain, but just switches the current consensus mechanism of ETH. The above applications, users, what to do, all habits remain the same.

So I said that the big merger is the transfer of power. Just like when Hong Kong returned to China in 1997, although the authorities have changed, the people are still the same, and the business is still the same.

5/
Theoretically, the best situation may be: before you go to bed, it is still PoW Ethereum. When you wake up, open Metamask and make a few transactions, you will find that it has actually become PoS. You don’t need to do any operations, and you don’t even have any perception .

6/
I won’t go into details about the benefits of PoS here. Simply put, it saves energy, improves performance, and makes it easier to determine the final state. I personally support it. The overall advantages outweigh the disadvantages.

7/
V God mentioned that in June, Ropsten, the largest testnet of Ethereum, will implement PoS.
This is different from before. Rop is the test network of his own son, the number one default network of Metamask, and the network that almost all Ethereum developers will use.
Once Ropsten is installed and there are no problems, then the mainnet launch will only take another 2-4 months.

8/
In addition, what everyone is more concerned about is when can the deposits of $stETH and $BETH be withdrawn? According to Vitalik’s roadmap, the main network will be merged first, and withdrawals will be enabled after a period of time after the merger.

9/
In addition, regarding the performance issue, I understand that pure PoS can only improve a little TPS, and don’t expect PoS to be much faster than PoW. On the contrary, the difficulty bomb started to explode in June. If there is no merger at the end of the year, the block interval may be pulled to 18-20 seconds, which will actually reduce the ETH TPS and increase the Gas price.

10/ As for performance expansion, it still depends on sharding. But sharding, to be honest, is really difficult to develop, and even difficult to design. Centralized sharding is easy, but when it comes to decentralization, the difficulty is several orders of magnitude higher.

No way, since ancient times, you can't have both fish and bear's paw.

11/
I still remember that at the beginning ETH said that it would have 1024 shards, which would directly increase the performance by 1000 times.
Later, the tone became softer, saying that we should try 64 shards first, and the target has been reduced by 16 times, hahaha.
Later, the community said that 64 pieces can’t be done, so let’s just engage in danksharding.

12/
Danksharding is relatively new, and it seems that there is no translated name, so just understand it as a new sharding route.
V God also reiterated this view today. In the past 64 shards, 64 committees were still needed. If Danksharding is used, these are no longer needed.

13/
At that time, combined with technologies such as PBS and crList, only one block proposer and one builder are needed to directly create a super-large shard (the goal has been reduced again, from 64 shards to 1 super-large shard ).
To use an imprecise metaphor - it is to make a super large block.

14/
How big is this super block? - 1MB.
After EIP1559, the current block of Ethereum averages tens of KB to one or two hundred KB. If Danksharding is used, the block size can be directly pulled to about 1MB in the first step, and the TPS of ETH can be increased to 5,000.
I personally feel that 5K is definitely enough, after all, it is only 15 now.

PS: I think this is different from the large blocks of BCH, that is, the introduction of PBS and crList technology, which can eliminate the centralization risks brought about by the large blocks.

15/
In the later stage, ETH strives to pull the block to 16MB (it may take several years), and then cooperates with Rollup technologies such as Layer 2. It is possible that the TPS of Ethereum can be increased to 80,000 transactions per second.
Then this data is adequate and sufficient, and now these web2s are of this order of magnitude.

16/
In addition, V God also reaffirmed the historical status of Rollup. Rollup is by no means a temporary measure before the expansion of Ethereum. It seems that this posture is going to coexist for a long time. And V god bitly praised the OP, saying that its management module design is better.

17/
Best, Vitalik also has a point of view that I think is advanced, that is, ETH is too dependent on DeFi now. But DeFi is OK in the bull market, and there may be many unexpected dangers in the bear market. God V also called $Luna a wave of 2333. Therefore, Ethereum needs more non-financial applications.

18/
However, I feel that this is a chicken-and-egg problem. Only financial applications and top-level artworks can bear the current gas fee of ETH. Others, such as social networking and games, I think are hanging. But if PoS upgrades smoothly and Danksharding progresses smoothly, then there is still hope.

19/
Looking forward to the merge!