chain?
What’s the argument for it not being on chain?
https://x.com/adriankbl/status/1808644935450607747?s=46
Oh the humanity. Any chance you can shoot yourself in the dick instead of posting here?
the arguments presented would essentially imply the billions of HCS transactions from atma are also not on-chain
I’m under de impression that mirror nodes are used for the heavy analysis and query work which you don’t want on the main nodes. They have to come with better explanation when they do these claims.
I read the following on docs Hedera: Hedera consensus nodes push new record files and signature files to the cloud storage provider – currently AWS S3 and Google File Storage are supported. Mirror nodes download these files, verify their signatures based on their hashes, and only then make them available to be processed. Just for understanding, if the record files which are pushed to cloud providers, get lost, can they be pushed again or are they lost forever?
I expect they can pushed again. Everyone can start a mirror node.
yes exactly, splitting the nodes helps optimize for tasks that each are great at. Does not mean they are not part of the network 🙂
There are a lot of redundancies in place - including cloud storage and other mirror nodes
So they can not be pushed again? If they can not be pushed again, the guy has a point
Someone would have to go and delete the record streams manually from GCP + AWS + all other redundancies
I think what he means, is the original data still on the mainnet and can be pushed again. So not pushed from a secondary source.
From my understanding the consensus nodes don't keep the record streams for too long
As we chatted about a couple of days ago state proofs can be added for a little extra spice… But with other networks it tends to just be hashes of things anyway, so the rebuttal would be “but ser your onchain data is just a hash of your thingie”
Yup exactly, these kinds of issues at this level are not unique to Hedera. Any "decentralized storage" or even ledger in general could lose data. The solutions are often just risk mitigation, e.g. more decentralization of nodes.
Make full sense- the whole point is not to clutter, but index and be able to link. Mirror nodes are for that and act as distributed replica as I understand that. Hedera has very good approach, it is fast and works well.
Обсуждают сегодня