So far my thoughts and understanding has been that if I have an amount of data that I want to store in SAFE if I am providing at least that amount from my node, it should wash. I “pay” to upload the data. So if I am providing 10 GB to SAFE then uploading and storing up to 10 GB should be covered.
Again, as I understand it, once the data is uploaded, it is in SAFE forever.
So if I then drop my sharing, I could go to another computer and access all that data as it is already there.
Ok, consider I am a malicious person (or threatened big data company or government) with lots of resources. I could assemble say a petabyte of storage and provide that to SAFE so as to to upload a petabyte of data. Once in SAFE, it is stored forever on the general participants provided storage.
So if I dropped my providing of storage, reconnected to safe with reinstalled computers and then provided a “new” petabyte of storage and then uploaded a “different” petabyte of data (I say different in order to defeat the deduplication aspects) and then kept repeating this process. Would this not result in lots of storage being used thereby taxing to good actors and having a lot of data/storage that is useless but causing the need to keep adding nodes and storage just to maintain?
The smaller the participant sample, the more this is a problem but could it not be used to cripple the startup?
This isn’t how the token economics work. The malicious actor would have to wait to receive payment for providing resources, and then use that payment to buy storage.
This is why pricing of storage is important, particularly at launch when the Network is young (I.e. less distributed). The price to upload should be fair but also steep enough to make it not worth someone’s time / too uneconomical to launch such an attack.
This is also another reason to moderate minting of Safe Network Tokens. To upload all that data, the malicious actors would need to get enough Safe Network Tokens to pay for the storage. Limited supply of Safe Network Tokens at launch will provide the Network time to organically grow and become stronger (I.e. more distributed).
Is networking/transport also part of the usage/provide/pay model?
And I still have the question as to how realtime communication would be done over SAFE. Like how would streaming a live call happen. On oldNet the sound would be broken down in chunks and sent from source to destination, then reassembled. On SAFE, is that done by file chunks? Would that be fast enough? Can live(near realtime enough to seem realtime)?
Your calculation is wrong. You miss the important fact that the network starts with 0 GB, and there are already millions of free terabytes available in the world. This means that the Safe Network can theoretically be doubled in less than 4 years. Of course here you have to include that Safe copies will also fight for these millions of free terabytes, so who knows what will actually happen…
couldn’t you just use the same formula as for the storage cost? network bandwidth, cpu efficiency, etc. gets also cheaper each day. And you end up with pennies for all old data on the network.
HI @IvanRainbolt This is all OK and by design. When you “pay” to store, however you pay it is stored forever. If you pay to store more than again it’s forever and OK. All new farmers etc. know they have to store all historical data and the network assures the farmers earn after they are shown to store all data, if not then other farmers will.
So the design sees such attacks as normal use and expects it, the more data the better and if folk want to pay to store junk, it’s also OK economy-wise, but useless for humanity. In any case, the network is fine with that.