Risk of Safecoin making the SAFE network fragile

I don’t know, we’ll see how much PUTs will cost.
I made some attempts to guess, but educated guessers on this forum have declined to speculate on that.

Secondly, you don’t know how much spare capacity will exist on the network. Maybe they’d need to strike at at peak hour (e.g. when it’s 93% full). If the network is 4 PB large, 7% would be 28 TB.

As the price of storage will change with demand (read: spare storage available), I wonder whether ‘filling’ the network is feasible at all. It may just make PUTs prohibitively expensive.

Ofc, ludicrously expensive PUTs wouldn’t be good for the network either, but it would deter these sort of attacks effectively. The storage cost would fall again as the network recovered (more storage came online).

Well that is what I said in one of those discussions where I was officially warned.

You can’t be both right and wrong.
If you are right, then my question is what does each Safecoin represent, when you can buy 38 of them today to upload 5 photos from a recent trip, and only to find that the network re-adjusted itself to respond to increased use, so now you can post only 3 photos instead of 5?

Either you’re right, and Safecoin will represent the cost of SAFE capacity at the margin, or wrong, and the attack could be funded by buying a sufficient number of coins in advance. If it’s the former and the network adjust quickly, the exchange of rate might fluctuate more than BTC vs. US$.

Have a look at the safecoin RFC, its in there

What will happen as the spare space drops because of the uploads then the sacrificial chunks (which are a measure of spare space when supplied space is 6xchunks stored or less) are being sacrificed the put cost algorithm raises the put cost very quickly. When the sacrificial chunks are gone (network still healthy == 4 chunks/chunk) the put cost is one coin per chunk put.

When there is plenty of available space there is 6 chunks per chunk stored. The sacrificial chunks represent 2 of those chunks, permanent chunks represent 2 and backup chunks represent 2. When sacrifivial chunks are sacrificed the put increases quickly

When Total_sacrificial = total_permanent the put cost is extremely low (can be 2^63 chunks put per coin)

chunks per coin drops proportionally to

TP / ( TP - TS )   as TS drops so does the number chunks put per coin drop.

i think the cost of storage will fluctuate against the cost of safecoin. How much that will cause safecoin to fluctuate against other currencies as a result is an interesting question.

We will certainly get more storage per safecoin as time goes on. Does this mean safecoin should fall in price by the same amount? I would say no - safecoin will float freely against the storage cost and other currencies.

If storage is the only thing that is bought with safecoin, then the relationship may be close, as the only value becomes storage utility. However, if storage is just one of many things bought with the currency, then there will be value beyond storage utility. This value will not evaporate when storage cost becomes prohibitive for a period.

All in my opinion, but the network will have to balance itself in terms of supply/demand for storage. This has to be done via the market and it is why safecoin will float against storage cost.

I meant in fiat, of course. Whether the cost of 1MB in Safecoin is 5 or 1 or 83, is not important, those are just numbers. The attacker would want to know how much it’d cost in fiat to put X TB’s of data on the network.

Only one cost is semi-fixed here, and that is the cost of disk space. It is tied to real world and gradually declines due to tech improvements.

But the float of Safecoin will grow quickly too, so who knows.

Semi-fixed being key. It will depends on the fluctuating supply/demand for storage on the network, which is obviously why we are discussing it on this thread.

Not really.

If cost goes to 1 chunk put per coin is a big difference to 2^63 chunks per coin.

Agreed that $$ cost is important.

But once you talk of filling up the network then you need some serious coin to do so, and the put algorithm serverly limits anyone’s ability to fill the network.

If the network is small then it follows that not much coin has yet been farmed so even if someone got all 400 million of the coin and starts filling the network then very soon become limited to 400TBytes more data they could upload.

And if the network is large and there are a lot more coin farmed and they got all the coin they would be limited to 4000TBytes once they started to fill the network.

400TBytes is 400-1000 farmers on small network

4000TBytes is << 10000 farmers on farmers on a large network

You see if you rapidly fill the network then you are very limited in how much you upload if you actually were able to start filling the network.

Who cares how much the coin cost in $$$ as long as there was someone who could buy nearly all the available coin, they still would find it VERY difficult to actually fill the network at a very rapid rate. You said 1-2 hours didn’t you.

Now if they did it over weeks limiting the rate to allow more space to be added then they get to upload a lot more, but that defeats filling the network to break.

1 Like

That is exactly why I have made this point.
If the cost goes to X, it can also drop to Y. So what does 1 Safecoin stand for, or represent?

Like any currency, safecoin will have to establish its place in the market. PUTs is on only one thing that it will buy. The price per x PUTs will be just one factor in the perceived value.

It’s not backed directly by anything, so “what it represents” is not like being able to say “one ounce of silver” etc. It represents the privacy, security and freedom to hold and share data in better ways, including exchanging value in various ways, for an infinite variety of things, including PUTs.

What does the US dollar stand for or represent? Currently it represents debt, elite control, war, etc., but it also represents something that most anybody will take in preference to other things.

I have some of the concerns you represent and agree they should be looked at and provided for. I just thing they are already provided for better than you think.

1 Like

I agree.
What I meant with my question about Safecoin was only that it doesn’t represent a fixed amount of disk space, so the argument that buying X amount of space would be prohibitively expensive is equally invalid as the argument that it would be next to free.

This attack would require close group diversity. Anything less could be prevented with a small algorithm added to one of the regulatory groups in the network. Close groups could be designed to, upon extreme conditions, collaborate for a complete network consensus to temporarily slow PUT speeds and identify nodes in XOR that are responsible for the largest amount of data input with low granularity (detail). Basically what day, and these operators >< (e.g more than 20 gigs or less than 100gigs. This is SAFE of course. Statistics need to be free of deanonymizing data.

In such an extreme case, a queue system can be temporarily instated by the network to prevent abuse that completely halts data injection. At this time, the network would prevent massive data dumps. Let’s say 4 gigabytes would for that time be the maximum at once for any given node. Anything above this would be re queued and resumed when given it’s next chance. If a script were written to automate the process, the attacker would still be subject to the queue. The network could at that point inform the users on the network that a shortage is being experienced. The network could leverage the event by informing the user of improved chances of farming safecoin. Incentivizing growth and regaining PUT ability without the need for queuing. Couple this with the rising cost per PUT and the network would be immune to this attack.

Safecoin is both a limiter and an enabler of data manipulation on the network. It’s data storage utility follows the very simple supply and demand algo. I’m not sure what being sought when the question “what does it represent” is asked. Technically/literally or philosophically? The philosophic has been done to death, technically we all know it to be a token. What it can provide us at any given moment is dependent on network status and the APP ecosystem. So my answer if I had to give one is the one I gave above with one addition. Access. Future applications will undoubtedly require payment for access to exclusive content or services.

As the network fills, the cost per chunk in Safecoin will tend to infinity. Therefore, it doesn’t really matter what the dollar price is - it will tend to infinity too.

The point is, the network is designed to increase the cost of storage to prevent the network reaching 100% capacity. The closer we get to 100% capacity, the more expensive storage will get… prohibitively so.

1 Like

Kinda like trying to travel at the speed of light. :yum:

3 Likes

But in one discussion - I can’t find it because so many topics have the same set of keywords - someone said that some parameter would have to adjust faster than it is now designed, because the network could run out capacity before the cost of PUT goes up.

Edit: found it. Read the whole thing or start from this link:

1 Like

I just read the last 50 or so messages in that thread and it seemed to be a debate between 2 none maidsafe staff, speculating on the implementation.

The white papers have stated that you will be able to buy storage tokens with Safecoin and that the price will change depending on supply/demand of storage. How quickly this ratio will change must presumably be quick enough to be useful and to avoid being gamed.

However, we don’t have all the implementation details yet. Perhaps people buying up loads of storage tokens will allow a sudden flood of PUTs which would be hard to manage. I am sure mechanisms to limit the impact of this could be brought into play though (perhaps the network should aim to always have enough space to cover these, for example).

Maybe it is a valid concern and I am downplaying it though. We can find out once we have something concrete to dissect.

1 Like

Yes it was, because - my interpretation - at the beginning of that topic it became clear that the current design has to be changed.

Last fall I estimated that the coin-related (economics) work won’t be completed by March and possibly even May; issues like these (there’s 3-4 of them, by my count) are still open and it will be interesting to see if they will be solved or merely worked around.
Right now I only remember 2, but I think there’s at least 1 more:

  • Farming rate algo (that we discuss here)
  • Rewards system (for example app authors creating fake requests to earn more Safecoin)
1 Like

It represents 1 coin in a total of 4.3 billion. That is the point. Once the cost of putting is so high that it is one chunk per coin then there are hard limits on how much you can upload NO matter the $$$ price.

And your example would drive the price to 1 chunk per coin extremely quickly according to the put cost algorithm in the RFC. 1 to 2 hours is not long enough for people to add storage so the network would fill.

2 Likes

And if you look back that is talking about [quote=“Seneca, post:24, topic:6182”]
What about this alternative:
[/quote]

not the RFC

There is a lot of material in that topic and most of it is about alternatives.

The RFC shows that as you in your example flood the network with puts to fill it in a couple of hours then the put cost cannot do anything else but go to 1 chunk per coin. Of course if you cannot upload fast enough to fill the network then it may not. But then you did not fill the network.

Remember that relies on your example of extremely ultra fast uploads

1 Like

Okay, you got it, my criminal mind has to come up with a way to address that:

  • an additional uptime and cost would be required to fill up the vaults of the attacking army
  • then the attack would commence by simultaneously PUTing data on the network and taking down their own vaults. That would double (or more) the change delta