Adding more thoughts:
Essentially, the value of zk/snarks on the scalability side actually comes from breaking the tradeoff between data consistency (security) and resource utilization efficiency (scalability), i.e. satisfying both sides at the same time, while maintaining decentralization. Yes, a solution to a well known classic issue - the “impossible triangle”.
Let’s walk through the underlying theory again: (under decentralization assumption)
- strong consensus ensures data consistency, as a cost it requires high replication of work, which requires high consumption of resource (which is the reason for high gas cost)
- to improve scalability and reduce cost, the ultimate way is to reduce the replication degree, but it can make consensus vulnerable and harm the data consistency.
- but with snark proof, the verification secures the data correctness in diverse work and the succinctness ensures low replication work. Therefore, it satisfies both high data consistency and high scalability at the same time.
zk rollup solutions made a valuable attempt, it satisfied both the scalability and data consistency, but it still requires consensus to maintains the consistency among sequencers in the future, therefore replication work is still inevitable, which requires high cost and may reduce scalability again. Otherwise, it has to sacrifices decentralization and keep the number of nodes small to stay low cost. In other word, the low gas cost of zk rollups mainly comes from sacrificing a certain degree of decentralization, the value of zk/snark is not fully realized. Rollups are still trapped by the “impossible triangle” (for op rollup it sacrifices certain level of security, but we won’t dive into the details)
So is the World Supercomputer a different thing?
Yes, it can fully release the value of zk/snark. Why?
One of the key points of World Supercomputer architecture is keeping the sequencer only in the consensus network (CSN), other than in functional networks, e.g. computation network (CPN) and storage network (STN) does not maintain their own sequencer. Therefore, CPN/STN can achieve scalability by focusing only on low replicated work, so that even when the number of nodes grows, the cost won’t increase.
Because we are not trying to reduce the workload of sequencer, which belongs to CSN, we are trying to reduce the workload of tasks that should be scalable.
In short, CSN == decentralization + data consistency, CPN/STN == decentralization + scalability, and use snarks to connect the data without losing any ‘angle’, to achieve WSC == decentralization + data consistency + scalability.
If we took a deeper look at the “impossible triangle”, decentralization actually is the strictest requirement, it requires ZERO centralization in any component of the whole architecture. Sacrificing decentralization to reduce cost is definitely a user-friendly solution, but it doesn’t help to solve “impossible triangle”. So the true challenge for Ethereum and even the web3 is how to bridge the scalability and data consistency UNDER the assumption of decentralization, the red line of web3.
Previously we thought this dilemma cannot be broke, but thanks to zk protocols for enabling it, the World Supercomputer architecture has become a promising theory to solve “impossible triangle”.
There is still a lot of work on both research and engineering sides, if you share the same ambition, let’s build the wsc together, the wsc needs your idea, comments and contributions. Thanks.