I could be wrong, but my understanding is that the network will price rewards solely in SNT and not in fiat. The fiat price will likely matter to up-loaders and farmers, but not to the network. Their behavior based on fiat (or any other consideration) will drive the network rewards in SNT. There are dimensions to this that will be really interesting to see play out. Each individual needs to make decisions on their own situation.
Okay so the human element of people seeing the fiat price/how much the network is offering in rewards drives a call to action to farm, and upon the network getting too many resources it balances by lowering the farming reward (and vice versa) till there it reaches its equilibrium but as the price is always fluctuating the network has to continuously and dynamically balance in real time. So as you say it somehow finds the right price indirectly by the network balancing its resource needs.
That is something kind of hard to wrap your head around when you really question it because of it being indirect. Thanks
I’m inclined to believe the fiat price will inform the store cost/reward b/c the fiat price is what will inform human behavior (e.g. how much they store or farm). How much humans store or farm will dictate the Network’s supply/demand economic calculations. So if the price of Safe sky rockets such that the reward is now extremely high, more people will farm, potentially making supply outstrip demand until the Network rebalances. Essentially this:
We’re assuming that the network doesn’t throttle new entrants for this to work properly. I think they were working on allowing new farmers to join more easily now instead of having to queue to join, so that’s good to hear if so (also just from a UX point of view).
Another factor on price downswings is that people will probably “hang in there” to a large degree in order to support the network because it isn’t great for the health of the network to shrink very much.
So hopefully “number go up” for the most part.
I was actually just thinking about that too. I was thinking with bitcoin the difficulty adjustment forces some miners out but on the Safe Network people are mostly running spare resources and so won’t mind earning less which means they will be willing to hang in there for quite some time. I think for a lot of people farming will be a very passive act.
What kind of impact does that have then? Does that essentially make the network grow more like I think you are suggesting? Seems like that could lead to some massive growth quickly once people catch wind.
It’d make for cheap uploads when the fiat price drops, which means the network will fill up until it reaches a store cost that is no longer considered cheap.
Also if there are rewards for gets, having more data on the network would help increase the farming profitability. And people will want to acquire some tokens to do some cheap storage, which would dampen the downward price action.
Its both and will be via human agent. Overall
- humans who buy tokens to upload will slow purchasing when price ($$$) is high.
- humans with tokens already will also slow down purchasing resources.
When farming rewards increase (lower spare space), overall humans will add storage resources.
When farming rewards decrease (higher spare space) overall humans will remove some storage space
This is a negative feedback loop system using human nature. While in some cases the trend will seem to be not followed, the negative feedback will bring it back into control.
You can think of it as a rubber band control. It can stretch but the further it is stretched the strong the pull to bring it back to reality (base price).
The base price will be determined somewhat by speculation together with what the market will bear. And from there the 2 interrelated negative feedback loops bring back the pricing to the “base price” when major jumps in price OR storage occurs.
For example the “base” price of SNT is determined to be 2$ then if
- spare storage starts to be scarce then the network raises store cost (Higher SNT to store)
- people slow buying SNT for buying resources
- people with SNT slow buying resources. Some start sell fearing the worse
- price $$$ in market starts dropping
- people buying SNT for buying resources reduce their slowing
- Scare resources increase network payout (SNT) for rewards
- people add more Nodes/space to get rewards
- network sees this and space no longer as scarce
- SNT store cost for resources reduce
- Rewards (SNT) reduce
- Thus the $$$ cost to store comes back to around what it was before
Now if cost $$$ of SNT is 10$ the change is only in the base SNT that the network will be using. The reason is again human nature.
When the price rises from 2$ to 10$ people find they can buy more resources per $ and then do so. This is an overall effect, some humans won’t but others store as much as their wallets allow. This causes the following
- more resources are bought.
- network storage begins to become more scarce
- farming rewards in SNT and $$$ rise dramatically. (SNT reward amt increases and the 2$->10$)
- more farmers come on line and with the 2$ to 10$ base price increase a lot more farmers come online
- This causes the SNT for buying to drop
- this causes a cycling until people slow buying extra resources since they run out of things to upload. At that point we reach a new much lower steady state for Store cost (SNT) to buy resources and rewards for farming.
Thus the effect is for a new base for store cost (SNT) to be made. I would think in human nature that the $$$ amount to store when SNT is 10$ would end up reasonably close to the $$$ amount to store when SNT was only 2$
In addition to resources, the rate of new nodes attempting to join the network and the rate of old nodes leaving the network is also a big indicator of fiat price. Although resources matter most, the actual node counts for inflow and outflow of operators may be a better indicator of fiat trends.
Exactly in line with AWS S3, 10x cheaper over 10 years (actually it’s less than 10 years for google drive, but I think it’s ballpark close enough to keep the simple 10|10 relation, and yes other tiers / timelines vary a lot with the ratio, but I’m gonna cherrypick )
Extrapolating this we end up with permanent storage being 5x the current cost to store for 1 year (calcs here).
So that should be around 300$ or 750 SNT per terabyte at today’s prices.
I am in error Ignore this post.
I took it was using real life example of how the cost per terabyte drops then charging 5 times the reward amount to store would be what we need to consider.
5 times the reward price for PUTs will over time cover the cost of permanent storage. So if GET rewards cover the physical costs of (Spare + dedicated) storage then storing is 5 times. Remember that spare resources is zero rated for that storage and dedicated storage has to compete on that grounds.
The only issue I see here is that storage cost is not necessarily correctly reflecting the rewards that will have been done (overall) during the lifetime of the Safe Network. @mav
For reference, pCloud sell lifetime storage at between 3.7 and around 10 times the annual rates, depending on whether they are on sale. The lifetime contracts seem to be permanently on sale, suggesting that at 10x annual rate they don’t shift many.
No I don’t mean to connect reward to this at all.
It’s just a way to look at current cloud price per year and try to equate it to a cloud price forever.
eg if I see a price of $1 per year for AWS then I’d expect it to be about $5 forever on AWS, and this allows me to make some comparison with whatever the price is for Safe Network. It’s just a rough heuristic to convert per year price to forever price.
Might be of importance to mention that big server/cloud companies have access to different hardware than the consumer market, which might give an unfair competitive advantage.
“The IBM TS2280 Tape Drive uses next-generation LTO Ultrium 8 cartridges to store up to 30 TB of compressed data per cartridge – twice the capacity of previous generations – with a ten-year TCO for data retention that is 86% lower than an all-disk solution and 66% lower than an all-cloud storage solution. Discover how to reduce your storage costs, simplify data management and enhance security and compliance.”
"Modern companies use tape to store their data when they need high-capacity data storage methods that offer high performance at a low cost. It is common for a magnetic tape data storage cartridge to have the capacity to store six terabytes of data at a third of the cost it would take to do so with a comparable hard drive that etches data into a disk or cells in a solid state drive. "
Please don’t shoot the messenger, I have no expertise in pro business storage, I just watch some tech news from time to time.
No worries. Storage can be divided into 2 (generally) primary and secondary (or tertiary). Primary is likely to be accessed a lot, secondary not so much. Tapes are used for tertiary storage mostly. So stuff like 7 years records and such like, where you run a report and the tapes can kick in (perhaps even be put in by a human).
For us that would be like archive nodes
If I understand the IBM info right it seems like the new tape storage solutions could be used by companies like Netflix or similar, with speeds above 350MB/s, closing in on last gen SSD’s.
But as you said it is probably for data which is stored long-term and not erased or re-written.
Given multiple people could be watching a video at any time, I doubt tapes would be a good idea. They are terrible at random access, which would be needed for streaming two different parts of the same video.
Tapes are a backup medium and as you say not good for random access. Even the old 8/9 track tapes used in mainframes of old were mainly for sequential reading/writing. And back then disk drives were not so fast and thus they had their place in bulk online storage. But with todays disk/ssd technology the tape drive is not useful as random access as days of old.
And of course once we talk of retrieving a chunk and being quick enough to be rewarded, well the tape drive would only be starting its seek and the chunk will be at the requesters machine already.
The speed quoted by IBM is sequential speeds and only useful to netflix when they use a hybrid system where a movie requested is read off tape and stored on disk. Then on a cache algorithm it would be removed from disk when space is needed for a new movie. This would save on expensive SSDs/drives because they do not need to store rarely viewed movies on them. It could see them requiring 1/2 the number of fast disks.
But for Safe nope. Safe is demanding 1MB random access with sub 100 millisecond response. Netflix does not require anywhere near that response time. You select the movie and by the time play is hit, it could be 10 or 20 seconds up to minutes since the viewer reads the description sometimes. In this time the tape drive is seeking the movie and starting the transfer to disk.