Correct. I feel zksync is a better engineered, cleaner solution, that does not introduce unnecessary complications.
It was pointed out above, that most storage node operators agree with me in preferring zksync, regardless of reasoning: Support for polygon was added, did not receive any traction, and was subsequently dropped. That’s the sufficient reason.
Why zksync was chosen in the beginning — polygon did not exist when decision was made. See Alexeys comment above.
So does zksync. And transaction rate is entirely irrelevant for storj usecase
Liquidity is a problem, yes, but that’s because there is no incentive to use zksync today. As long as storj is paying gas — why should SNO jump through hoops with any L2 solution?
I myself started with zksync and then switched back to L1.