The Economics of safecoin

I just posted an article on the economics of safecoin on the MaidSafe blog:
The Economics of Safecoin | by MaidSafe | safenetwork | Medium It is an attempt to carry forward much of the discussion we have been having on this topic in the forums and mailing list. I would really appreciate if you could check it out so that we can discuss concerns with the approach, what gaps still need to be filled, potential enhancements…etc…

Many thanks.


Thanks, Nick. It’s a great coverage of the topic overall, and a great place to start the dialog.

(1) I think the Tragedy of the Commons(TOC) is much less a point to be considered than malicious attack, especially in the beginning. But either way I think the solution is the same. We’ve discussed aspects of this on other threads. Make access free to anyone who can prove (via a Captcha or the like) that they are human. The free access comes with a relatively small storage and a set of basic, though very useful, features but not the premium, resource intensive ones. This prevents machines from creating a billion accounts and swamping the network, and reduces the individuals ability to do the same to an insignificant level. Also there are all sorts of reasons one might need to have a second or third account. @dirvine proposed that the minimum then would increase on that account over time to the network average. I’m not sure if that’s desirable or not but at least the entry curve for malicious attack would be reduced.

From this angle everyone would be able to adopt usage for free and thus really build fast, without risk. Contributing resources either by farming or buying safecoin, or by building apps, could then upgrade to much higher functionality, but rewarding the maintenance of the network.

(2) The value of safecoin will be its scarcity, anonymity, speed and certainty of transmission, etc. If tied too closely to data storage limits, app usage, etc., it will distort its value curve and perhaps even harm the network balance. For instance, if the free access is functionality is set too low, requiring safecoin purchase to make it useful to most people, this could become a problem if the value of safecoin goes through the roof. This would harm the network. On the other hand if that balance is right, it will incentivize builders, core developers and farmers to really expand the network, because the safecoin reward is so attractive.

So striking the right balance between 1 and 2 is the main decision I see to be made.

(3) I do think that other tokens/currencies can be built on top of the SAFE network, but safecoin is at the core of it and will almost certainly be dominant.


Regarding the article,
I like the freemium model, if done correctly. But it is very difficult to anticipate how it will work in practice before launch. The specifics of a freemium model should consider the following questions.

Q. Who is in competition with the SAFE Network?
A. Initially, cloud storage services, and other decentralized networks.

Q. What advantage would the SAFE Network offer over these competitors?
A. Cheaper online storage, combined with decentralized internet, more security, more anonymity, plus a native crypto currency (safecoin).
@nicklambert We need a competitor comparison table/chart.

Q. If competitors, private companies, can match the same features of the SAFE Network, why would people prefer SAFE over them?
A. The SAFE Network is not owned by any individual person or company. Rather, it owns itself and is shared by the community. If a fork/clone was made, the SAFE Network still has the first mover advantage. If the fork greatly improves the functionality and features, then it deserves to take the reigns.

Q. If better routing and encryption is developed, will the SAFE Network be able to evolve?
A. Often times, when a building is constructed, the “foundation” is made permanent, while the applications on top of it can change easily. I don’t know.

As many have stated, if we give away too much FREE storage, the system goes broke. If we give a away too little FREE storage, then farmers have very little to farm. This is a conundrum.

The Network Average calculates what is being stored. If the network starts with ZERO data being stored, won’t that mean the Network Average is zero?
EDIT: I’m assuming the first farm to start the network will already have some data stored, similar to how Bitcoin had the first bitcoin mined. That should answer my own question.

Alternative Freemium Model

  1. Make the NSL (Network Storage Limit) based on Farm Rank. This partially emulates the POR system without slowing down the Network. Users who contribute a lot of resources should gain more storage space. If we assume most of our users will be farmers in the beginning, we should let them store as much as possible! I doubt farmers would spend their safecoin on storage.

  2. Because of #1 above, the farmers will set what the Network Average will be. Now we have an effective starting point. Farmers farm each others data, and also farm free users data as well. All accounts start at Farm Rank 0, with the ability to increase their NSL by increasing their Rank.

  3. The Network also charges safecoin to increase storage above the NSL. This is the premium service for large data storage.

Farm Ranking System
Rank 0 = ADS x 0.01 (Users is not contributing any resources)
Rank 1 = ADS x 0.5 (Small Farms)
Rank 2 = ADS (Average Data Stored) (Common Farms)
Rank 3 = ADS x 2 (Reliable Large Farm)
Rank 4 = ADS x 4 (Specialized Super Farm, Vital for the Network)

Example… Assume the Average Data stored was 500Gb.
Rank 0 = 5Gb (Users is not contributing any resources)
Rank 1 = 250Gb
Rank 2 = 500Gb
Rank 3 = 1000Gb
Rank 4 = 2000Gb (Specialized Farmer, Vital for the Network)

Google Drive is offering 15Gb free.
We offer 5Gb free with the ability to leap to 250Gb with a small farm. :wink:
If user has a common farm, then 500Gb :sunny:
A reliable large farm, 1000Gb :smile:
A super farm, 2000Gb :smiley:

EDIT: @happybeing I have reconsidered Farm Rank 4, and agree unlimited was too generous. Changed Farm Rank 4 to x 4. Thanks for pointing it out.

1 Like

Great SAFE article :slight_smile:

Main point is adoption by large amount of end users not farmers
And we need to build good, reliable, fast, pretty, apps especially for smartphones but also for web (site plugins), and desktop (science, big data). One killer app can bring swarm of users.
Smaller companies can also benefit from SAFE. Cheep alternative solution to storage, mail, team collaboration, …

1 Like

I have to partially disagree.

Farmers and Builders are also the End-users. Without them, there would be no Network, no safecoins coming into existence. We need everyone to make it work. :smile:

I like this. It’s simple and easy to understand. I would note though that it would be good to have the NSL, NR, ADS and such be displayed somewhere clearly for the user in a kind of HUD somewhere as they would be important stats to keep track of. Also I’m curious, why would one lose NSL if maidsafe just makes references for de-duplicated files? Like if I’m uploading 50-100gb of music and 90% of that ends up getting deduplicated because it’s already been uploaded by various other people then shouldn’t my NSL only be deducted for the remaining 10%? I mean if deduplification works like the way I’m describing it wouldn’t that mean mining would require one add original data in order to mine safecoin?

@dyamanaka I love your posts! Though I’m thinking like @magussilikonski here.

@dyamanaka I think you are making assumptions, then setting network parameters up that will make them come true.

Let’s imagine the network we want, then set the incentives in favour of this vision. If it’s not working, then perhaps we can, or rather the network can, adjust the parameters to get a balance between what we want and what actually works.

@dirvine had some inspiring visions that I think massive server farms could undermine if they are over encouraged/incentivised, so it could be a tragedy if we inadvertently create this instead of something that meets our goals and values.

I agree. Your NSL should only be reduced by the 10% that is new content. Because your 90% is already uploaded.

Farming is based on GET requests. Say 1 movie is about 1GB in HD quality. That file is shattered into 1Mb chunks and made into 4 copies. So you have 1000Mb x 4 = 4000 Chunks spread out on the Network. Just from calling that 1 movie to stream online, you initiate 1000 GET requests from a pool of 4000 farms. Farmers will be overjoyed you are watching that movie. Now add everyone else who is also requesting that 1 file. Example 100 people watching the same movie = 100 x 1000 = 100,000 GET requests from a pool of 4000 farms. Farming should be buzzing like crazy as the network gets larger, even with de-duplication.

This is the reason why we believe download will be very fast. Even with latency, you might only be limited by your own download bandwith. That is the theory anyway. We will have to see how it goes on testnet3.


I have to make assumptions. If I had a crystal ball, it would be easy.
My parameters, which can be adjusted, are generally based on sustainability first, then mass adoption. I’ve seen many start-ups fail because the model was not sustainable.

I agree about the dangers of massive server farms, which is why I’m more than happy to hear any alternatives, even ones that go against my solution. This is a community effort. :smile:

1 Like

Then first let’s get some consensus on what we would like, before limiting and containing this with assumptions (many of which are unconscious) about what will work.

We’re creating a new paradigm.

As an ideal, I suggest we’d be happy if the network was as diverse as possible, including all manner of users, devices, capabilities. As inclusive as possible, and self-regulating in ways that make the network sustainable while maintaining this inclusiveness.

This though requires agreement on other values, not just inclusiveness, since just being sustainable is not good unless it is delivering on those values. Survival without those values is bad, and such a network must be allowed to fail, or perhaps programmed to fail.

Some to start, which can then be debated, fleshed out and added to or pruned: Anonymity, privacy, accessibility, independence, equality, human rights, justice, sustainable ecology, unity and diversity, creativity, humanity.

There’s a start. If people agree it’s a discussion with having, we’ll need a new thread, so like if you think we need a thread to discuss the values we would like SAFE to support.


Thanks for commenting John, much appreciated!

I think on your point 2, I don’t think it is possible to disassociate safecoin use on the network, IMO you either use it or you don’t. I think we would need to fix the rate between safecoin and how much storage, for example, it buys so that as the unit price of the coin goes up it does not push up the cost of storage. So, if safecoin is worth $1 and buys 10GB then rises to $10 each it will buy 100GB.


@happybeing it seems like that is a different conversation. Yes, we should open a new thread regarding SAFE Project Values.

I was responding to the request from the OP. I offered an alternative because the economics, IMO, didn’t seem sustainable in the long run. Here’s why…

Future scenario,
In 10 years, The SAFE Network is very popular. The total amount of safecoin has reached its cap. Unless there is a constant recycling of safecoin by the Network, millions of farmers will have very little safecoin to farm. This could lead to a net reduction in farmers, which leads to a net reduction in total storage space available.

Now add the average storage space allowed to a swarm of free users. An imbalance could occur where free users are uploading more data than the Network can distribute to farmers. This is a unintentional DOS attack. While it does not affect data already stored, it will affect users who are writing new data.

Example: posting on your safebook page, sorry no space available. Sending an email to your friend, sorry no space available. This is very disruptive and leads down a vicious cycle where more users quit because their experience is turning negative.

More users will add storage to the network and increase the space available. But the Network is now sputtering like an old car with no gas. This is a scenario I hope never happens. And it is the reason I am trying to make sure the Network is able to manage such an event.

The reason why people put limits on scarce resources is not because they want too. It is because the resource itself is limited. It won’t matter how much safecoin is in existence. Sold out means sold out.

We “hope” people will just contribute resources to the Network to keep it running. I am very optimistic about that. Unfortunately, the analyst in me thinks we should ensure people keep contributing resources while limiting those who do not. A pure POR system effectively manages the above scenario which is why I modeled my solution to emulate it.

If this community believes people will keep adding more storage than they use. Then go forward with the plan mentioned in the article. I will not be offended and will continue to support this project as best I can.

(note: this ended up being a bit of working it out as I wrote, so see if my logic holds)

Of course, safecoin is always and forever associated with the network and will be used in transacting business on the network. What I’m talking about is avoiding linking the VALUE of safecoin to network resources directly. Your example goes along the lines I’m talking about.

The problem is “what is the yardstick of value?” The US dollar might be a reference point at the moment, but it also could go bozo valuewise, so the valuation on the SAFE network can’t be denominated in any external currency.

Working out that mechanism for determining the price, in safecoin, is the challenge. A value of one safecoin is very low at the network launch but hopefully will become way, way too much very soon.

The only reason I can see, at the moment at least, for having any direct price link between safecoin and any specific network quantity is to facilitate a low-end free access with a buy up to NSL or above. The only reason for this is to protect the network from being stillborn in bootup and early life due to malicious attack, and to continue to ensure more resourse is provided than used throughout the network’s existence.

I’m more and more liking @dirvine 's idea of slowly but automatically increasing the initial low storage access, up to the NSL over time, though I’m not sure whether that would solve the abuse problem, or just delay it.

Taking a further look, let’s step back to the base value that the network is built upon and denominate that: proof of resource (POR). That is, essentially, the gold standard of the network. Safecoin value will have to vary in terms of how much resource it will buy. This is complicated by the fact that POR is not just gigabytes of storage, but bandwidth and processor capacity, as well.

I don’t know how sophisticated the POR algorithym is or will become but POR is the very heart and soul of the network’s continued existence. Consumer adoption is also critical, but even that rides on resources.

So, I’m now thinking that a minimal account for anyone and everyone who can prove humanness (say, 5 gb). Beyond that it’s pure POR, either via equipment resources, app usage, etc. Otherwise, the tie between safecoin and POR will handycap the economics of both.

This leads me to the thought that the proposed POR token is going to have to be implimented from the start. The POR token would be the currency of the network itself. It expresses the actual input/output values of maintaining a functioning network. Safecoin, because is has a digital scarcity and functionality inherent in the network but is not tied directly to its functioning, could then ride a separate arc of valuation in relation to both POR token and anything else that one might find it useful to transact for. If you need more resource than the network provides free, you obtain POR token at market price of safecoin, or bitcoin or the dollar or whatever someone wants to sell it to you for, and you thus purchase that resource.

The fact that safecoin is distributed via farming/app usage which also come from contributing to the network will help encourage network infrastructure. But the value of it will float more freely and have the chance to appreciate more because of the lack of a direct tie to the network.

I really don’t know what the difficulties or other aspects of implementation are, but I’m starting to think that POR token can’t be avoided. Safecoin itself has already come into 10% quasi-existence with no network there yet. Its value is already decoupled, actually. It has a tremendous potential for upward valuation and an enormous range of possible uses. POR token may have its own value fluctuation and may become its own currency to a certain degree, but its value will be tied very directly to the resources of the network itself. That’s the currency the network directly requires for acquisition of resources that you are not directly providing.

[edit to add]

The characteristics of the POR token could also be quite different than that of safecoin, and could be designed to foster the indefinite upkeep of the network. I’m not sure exactly how that should be done. Perhaps a perpetual spend/burn cycle, etc. Maybe when a POR token is cashed in, it is burned and new resources must be added to create more. This would allow for the fact that computers have a limited life. Resource is always going to have to be added, just to maintain an equalibrium, much less expansion. The dynamics of this will require some real thought, but the purpose is very different from that of safecoin.

POR could also be bestowed as donations to those needing the resources. For example someone needs a particularly large chunk of resource for a research project but doesn’t have funding to buy on the market. Rather than appealing for safecoin, dollars or other currency, they could appeal for donation of POR directly from others who had surplus. Those providing excess resource would also be earning safecoin (at least for a good while to come), so would likely be much more inclined to part with POR on a liberal basis. If bots or malicious people did get accounts, they would still have to run the gauntlet of communicating with real people in an appeal for more resource, providing another layer of security that the resources aren’t being gathered maliciously. Those who are actually earning the resources are able to voluntarily and selectively bestow them rather than having the network do it at random, possibly without need.

Do my points at all comport with the reality of the technical end?

1 Like

I agree David. I think we do need to be thinking about a plan B. I believe the solution you offered is potentially feasible and something to chuck into the ring with any other suggestions.


@dyamanaka I see you tweaked your thoughts a little and acknowledge me for that. I wasn’t objecting to your thinking about these scenarios or you suggesting solutions though. As I said, I love your posts and I’m very glad we have you doing this work, which I agree is crucial to creating a system that works.

So what was I saying!? I think you did get it, but want to be sure.

I am not saying don’t think about this or post as you have. What I am saying is that we also need to be careful in these plans, that we don’t encourage scenarios that aren’t what we ideally would like, unless that’s the only way. And that if we start with a system with incentives that are designed to create what we’d ideally like, well we just might get it!

…and following on from that suggestion, I’m saying we also need to agree what the system is we do want - ideally - to create. Which ultimately comes down to what values do we want to support. And yes, I think that goes in a separate thread, but one which is used to guide and sanity-check threads such as this - where the nuts and bolts are being designed.

Hope that’s clear, and that you don’t feel I was criticising you or your proposal.

Only three people have indicated support for that proposal, four if you are also (not stated), so I’m not sure its going to be a worthwhile discussion - five people (including me) working out the values for the entire project?

1 Like

I don’t think you were criticizing at all. Values are very important. While the Network is mechanical, humans are social. I recognize what you’re saying and I get it. :smile:

It’s hard to express one self with text because people cannot see facial expressions or physical posture. And people often refer to my writing style as… strong? Anyway, it’s all good.

1 Like

I am trying to figure out some math based only on SAFEcoin but I don’t understand exactly how SAFEcoin is generated and how often? Is it just by usage with no time limit or its something else?
I am sure market will dictate the price of SAFEcoin in relation to the resources it can buy. (similar to mining contracts)

… still thinking … more to come … :slight_smile:

I am reading Whitepaper but just supplying of resources is not generating coins, only usage. Right?

We have to separate end users from SAFE providers.
End user just need space or computing power and they will not be bother to provide POR because for that you need always on PC (spending electricity and you also need PC, smartphone is not ideal solution for POR)
Farmer can also be end user but his main task is to provide reliable (always on) resources
Developers could provide POR but they shouldn’t be bothered with that if they don’t want.
Companies can be interested in POR as that can lower their monthly bill for using SAFE, but they still need some IT to manage that.

That’s correct Magus, safecoins are farmed when data is retrieved by the network from a node (gets). The network also stores data at random on the network so it is very difficult to predict how often coins will be generated, once we start to put the test networks in place we should have a bit more information to go on. The first network should be up and running in the next couple of weeks.

1 Like

The end user doesn’t need to provide an always on device, they can provide POR on a computer that is switched on intermittently, they just won’t gain as good a ranking as an always on computer.

1 Like