They should be more safe, easier to access and no more checking good condition of HDD.
Same reason the SAFE network will be better then servers. Sure an external HDD or some kind of cloud storage is better then no back up. What if there is a fire at your house or the data center though? Then its lost. With the SAFE network there won’t be one location that can get hit by lightning and destroy everything. To lose you data on the safe network many specific places would have to be hit by lighting (and all at once before it had time to recover)
I have 1-2 TB I’d like to upload. I think much I’d just upload as published data. I miss the file sharing of the past where people would share their whole hard-drive or at least make a public section of their drive to share. Now all that data is mostly locked up (perhaps deleted even). SAFE could bring so many files out of hiding. For example I have lots of tracker module files laying around and if I look at the various demoscene archive sites, I see that many are not on any websites now.
I wonder if I could still get data out of my old QIC-80 tapes and floppy disks to upload.
Just don’t upload the whole tape once maybe it will overheat?
I plan on holding every coin I earn for ~ 3 to 5 years. If fluctuations show no major rise in value I will use them to help migrate large data from the clear. Personal data is about 2TB.
For me, it really depends on where the pricing settles down at. I don’t want to give up more than ~1000 Safecoin, so, whatever that would translate to in storage. I’m hoping that will get me at least 5TB.
This desired price point has some interesting implications
5 TB is 5242880 chunks
If that costs 1000 safecoin it’s a price of 0.000190735 (=1000/5242880) safecoin per PUT, which is not possible with the temporary StoreCost algorithm proposed in RFC-0057, which has a minimum of around 0.005 safecoin per PUT.
Using AWS S3 pricing of $0.023 per GB
5 TB would cost $117.76 to put on S3
If this were equivalent to 1000 safecoin that gives a price of 0.11776 (=117.76/1000) USD/safecoin
I’m gonna guess 5 TB for 1000 safecoin won’t be possible in testnets because of the temporary StoreCost algorithm.
I’m gonna guess the USD price will be higher than 0.1176 usd/safecoin at launch, so will probably be more expensive than S3.
Sorry Let’s hope storecost gets a revision which would fix both pricing problems.
Thanks for giving the numbers though, it’s really fascinating to see pricing expectations/desires laid out clearly like this.
why dont they make storage very cheap from the start? so people can upload like 5TB for a 1USD and then pay vaults with the money that people paid to upload those TBs
There is plan that after revision of test SafeCoin they will change StoreCost algorithm. It is not hard to add 1/1000 multiplier to even proposal one as it is RFC only.
Maidsafe don’t sell cloud storage, so they don’t control the price. The price will fluctuate according to supply and demand (storage provided by vaults and usage of disk space).
Also, dont expect this price level. Uploading 5TB will use more storage than that because of data duplication. 1USD won’t be an enough incentive for vault operators.
But he said, he do not want use more than 1000 SafeCoin which might be worth $1000+ soon. The RFC 57 has low StoreCost only with no full nodes. So with 100k+ nodes and 0 full nodes you have StoreCost bellow 0,000 01.
Yes, this was my concern. At launch, I don’t think it is unfeasible that we see $1/coin. In fact, I think that estimate is pretty conservative. If we can’t get 5TB from ~$1000/USD, that is a problem that needs to be addressed. Trying to sell people on something more than $200/TB, even for permanent storage (aka, as long as the network exists), would be quite difficult for the mainstream, and I believe would be seen as a network failure.
If every test account will get 1M testSafeCoins than the StoreCost will be not issue. So what is plan about this ?
The test safecoins issued in the last test network, will become real safecoin…
So, just giving away millions of safecoin is not going to be great for its fiat price.
5k users on the forum, that would even be more than max supply of safecoin.
On the other hand, artificially setting store cost extremely low in the test networks, makes sure that a lot of users’ data can be uploaded, without them needing to spend precious safecoin in the very beginning. (Since it seems store cost will be higher initially, the first users - everyone at this forum - will spend a lot of safecoin on their initial data. As we know this, we might wait with uploads until store costs drops, but that might give a very slow rate of upload in the new network.)
The effects of setting store cost very low in test networks, seem to be:
- The test nets will see a lot of participants and a lot of data traffic, as everyone wants to take the opportunity to upload their data cheaply.
- It will attract users to the testnets - to take advantage of the above, making it bigger and safer when it becomes the live network.
Drawback is that it won’t actually test the real farming algo.
Yes you’re right that a lot of nodes with spare space would make the storecost cheap, but rfc-0057 says storecost calcs are “related to the numbers of nodes within a single section”. 100k+ nodes in a single section isn’t possible so likewise low storecost isn’t possible. My suspicion is that storecost will end up including the full network size somehow so the situation you describe to achieve low storecost will probably become real life at some point (just my guess though).
Makes sense. Consider TSC a pass to make it possible to populate the network without touching any SC or affecting its corresponding value. TSC will just allow us to witness the basic premise of spending a token to gain network resources. A volitile picture of the ebb and flow of transactions.
The beauty is that TSC will allow there to be some life on the network for launch appeal. Otherwise v1.0 will be a ghost town for a while which will imo slow participation in regards to the greater the public. Financial ramifications should minimal if non existent since farmers earn for GETs not PUTs.
So it sounds good to meh!
But since Fleming we will be able to run vault and store chunks with test SafeCoins based on SAFE Network Dev Update - June 13, 2019
So how you can get these test SafeCoins ? I do not think, that these will be once real SafeCoins.
In Fleming it is likely these will all be killed anyway. Fleming will not upgrade so will lose data on update to Maxwell (inc safecoin). Later though it is a different story
I guess it should be dirty cheap in the testnet or even free to upload just so we can get a jump start in the network and if it changes to a real network as it is then put cost in safecoins so that gets can be paid. so testnet cheap or free to upload to get jumpstart and then when normal safecoins are issues start with a random (kinda “expensive” so that the first real gets get paid good) and then the algo will make the price automaticly where it should be?!
Yeah sorry, not Fleming. But beyond Maxwell, any testnet could be “the one”.
Doesn’t need to be though (a bit depending on final implementation, but as it seems now), since it’s issued from the network supply. So the PUT cost does not need to cover the GET reward at all times. (After reaching stable state at desired 50 % issued, it would be needed to cover it on average, to stay there). The effect of not covering is that there is a net issuance from network. If we were to cover from start (or even exceed) then non-issued coins would stay at >85%.