Exactly, ICP is more of a cloud than a blockchain since we don’t get public records of what is going on behind the scenes. While BTC, ETH, Solana etc, you can see it all. CycleOps is a great tool, when you are the controller of the canisters, which in my case for the NFTs I was not.
Valve is an +8B dollar company, it has some killer games with stable sources of revenue that makes it profitable, they are community oriented like Dota2 and CS:GO among others with tournaments funded by the players.
It’s reasonable to say that is not eternal but but to think they will one day just shut down the servers, I don’t think so, and a great example is Dota Underlords, they halted the development (for 10 years at or until further notice) but still have thousands of daily active users(almost insta-queue even on ranked games) with servers up and running.
I wish we had on ICP at least one project that generates enough revenue to pay its developers outside grants and prove that is possible to thrive on Web3.
Going back to the cycles point, I just think that the freezing canister threshold should actually be a safeguard method to prevent it from being deleted, not a 30-day death sentence because it’s supposed to still have some/few cycles.
Or the canister only gets deleted when it completely runs out of cycles?
That’s just what it is. If you set a freezing threshold of 30 days, then the moment the canister can only pay for 30 more days of data retention, then it stops responding to function calls. You (or anyone actually, canisters can’t refuse top-ups) then have 30 days of leeway to top it back up. If the canister is not topped up, then it gets uninstalled (data deleted, canister still exists). If backups exist, it is still possible to revive the canister at that point. Only after a long time (10 years?) is the empty canister actually deleted.
IIUC you were under the impression that the freezing threshold says that once a canister runs out of cycles then it will be deleted <freezing threshold> days later. That often is how it works in practice because canisters often only drop to such a low cycles balance because the devs/owners actually have abandoned any upkeep of the project. Active projects usually don’t want to get to that point in the first place.
See It’s good to learn something new every day, I had no idea you could increase the freezing threshold of a canister, I just read the documentation and it’s pretty straightforward:
I was under the impression that you couldn’t do this, and the freezing threshold was set to 1 year by default and then reduced to 30-days, but these settings entirely solve the cycles management issue. (And no, I wouldn’t set a freezing threshold of 10 years)
The problem I had with my NFT collection is that I’m not a controller, I got a deal with Crowdfund NFT, now Funded and Entrepot, I wasn’t even topping cycles until a few years ago that I saw it was constantly freezing.
I know that’s not your problem and shouldn’t be mine either but it’s something that I’m dealing with from my own pocket since 2022. Yes I know there are tons of devs/owners that have abandoned their projects, and we can’t expect them to live off grants or charity forever if they can’t even make dapps sustainable.
Ugh in that case I don’t think anything can be considered permanent on ICP including any of my funds I deposit anywhere. Banks in real life have government guarantees (FDIC) for customers to be protected should the bank go down. No dapp on ICP has that both due to lack of higher-level insurance and due to lack of recovery mechanisms. So if I were to use anything it’d be with the knowledge that any of those canisters can just go down within the next year since not much has been proven yet as long-term sustainable.
Now this here intrigues me. You’re saying that if e.g. KongSwap went bust and ran out of cycles, I with my lone little account can fund enough cycles for it to keep going long enough to take my coins out? And perhaps automate this with my own canister that fires based on such an event?
That sounds good.
I know also that Dfinity plans to make query calls cost cycles as well sometime down the line. That’ll definitely shorten canister lifetimes if they’re idle, so I can only hope that the app developers can maintain their cycles easily and for the long term > 10 years.