Safenetwork sustainability concerns - Bandwidth has an ongoing cost however Safenetwork is a pay once, benefit forever model

See, here is the thing, with it’s current economic model it’ll SOLELY depend on two things from my understanding:

  1. How much NEW data is stored and how much safecoins people paid to store those new data. So If this month we get not much people storing new data on the network, the 100GB you donated to the network will make less money. If this month we get a surge of people uploading massive amounts of new data and paying a lot of safecoins for it to be on the safe network, then you will get paid more.

  2. How many other farmers are on there donating space. The more space donated, the less every farmer receives.

But regarding what you said regarding “pay extra or have earned money”, well, if you have used 100GB right now it’s a one time payment, not a recurring payment, so it’ll be a one time fee regardless when you stored data on there. Regardless if you then donate 100GB or not. If you meant operating cost of running the hard disks and giving 100GB to the network, i wouldn’t know, but those two factors i outlined above determines your profits, which also depends on other farmers, but the whole economy will eventually depend on how much people pay to store NEW data onto the network, so all the farmers profits will probably depend on that too, hence the number of farmers. So basically, No.2 is pretty much dependent on No.1 which makes the whole network somewhat dependant on No.1, which i feel may be a problem.

So i was suggesting to make a special rule where first, there will pay recurring fees for storage, and second, if you donate your space proportionately to the space that you have used up in the safenetwork, you only have to pay the one time fee and no recurring fees whatsoever.

I am not sure you really meant this, or did you ?

This is what I mean though, I suspect farm early (in a growing network) will be donating less than you pay for (so quids in really).

We are not seeing the same thing here. The costs to Put data depend on the whole amount of data that has ever been stored on the network. Farmers earn above this amount. So to farm you will need to look after all existing data at a minimum to continue on the network handing new data.

Here I see your point more, if there are only browsers and no store requirements then the network would (possibly) fail. I think this is unlikely, but I am also not sure it means definite failure, take skype for instance it never paid “farmers” (the nodes that routed calls etc.) but the network succeeded and made money, not from farmers but form clients who did not necessarily “farm”. I mention this as a counter example to your thought experiment.

So remembering SAFE has no concept of time, could you reword your idea a bit to allow us to keep looking at it. I think we are close but not quite on the same page. If you can do analysis without time then great. Otherwise the network needs a time mechanism, A sync mechanism was promoted previously in the dev forum though, but not flushed out fully.

3 Likes

I smile at this too. Just bought a new phone with good storage. Paid over 100$ more to get one with more storage. Then I bought a SD card for it (64GB Fast) and it cost me too.

From estimates it would have been cheaper to have that 64GB on SAFE than paying that much for the card.

So my *free* phone storage cost me a pretty penny and if something happened to my phone or when in 2 years I do it all over again I pay that again and in 2 years again and in 2 years again.

Its paying in extremely small increments and not paying again to restore those photos etc. You buy a phone or hard disk to store the photos then you have paid a BIG amount and then when the drive is too small or dies or the phone needs replacing then you pay it all over again. So non-SAFE storage is an ongoing PAYING for storage, maybe ongoing in terms of years, but take it from someone who has been buying memory & storage since the 70’s the price certainly adds up. And don’t talk to me about the losses and the additional costs of backups all the time.

Oh and the data I stored 5 years ago resides on less than 1/10 of the new drives. So keeping older data on a viable working SAFE will require less and less relative disk space. For instance all the family vids I took 15 years ago fit into the same space a 1hr 4K vid I took a few weeks ago takes up. The data stored in one year of operation will fit in 1/10 of the storage in 5 years time.


My thoughts on any type of rental system is that less people are likely to pay for this storage than use their own disks, What if forget that account where I keep my really old stuff that I don’t want to lose. Like my deceased parent’s photos that I look at only every few years.

What about that library that puts things up for people to use but rarely looks at the account themselves. Its an APP they developed that calalogues the material and stores it in the “library”. Too easy to forget to pay the rental and its gone.

What of personal libraries of technical material that is public and so many engineers & technicians use on a daily basis for their work. And the person who set it up and added to it dies. His family and friends most likely don’t even realise he did this on a special account and the tech library dies. And some of it was extremely rare to find stuff. Gone. There is stuff from the 70’s that I cannot find on the web anywhere including the POC stuff I did in the seventies that everyone uses on the web today.

You do realise that the charging model that is currently in the RFC is dynamic and charges the required amount to keep the network viable. While it may need tweaking the principle is to charge the minimum amount and yet keep the network viable under current conditions and it dynamically changes as the network storage changes.

EDIT: I wrote this before I realised there had been so many responses below.

4 Likes

Right now, how many people are merely using browsers and phones to access websites, every day? How much data are being downloaded daily? vses how much NEW data are being uploaded onto the website servers etc?

So from the image here, we can see that just WIRELESS data is used 18264GB EVERY minute. And netflix users stream 80,000 hours of video, while ONLY 400 hours of NEW video are being uploaded. So, lets think for a bit. First, Youtube videos are FREE to upload, not only that, it PAYS the uploader, with safenetwork videos have to be PAID to be uploaded, so if you do think about that, lets just assume the amount of videos uploaded to a site similar to YouTube on the safenetwork is halved because of this, so 200 hours of new video are being uploaded to a site on the safenetwork that’s similar to YouTube, while 80k hours of video are being streamed every minute, and popular ISPs right now have a data cap per month and then charge $10 for every 50GB hence $0.2 per GB on average ( Data cap analysis found almost 200 ISPs imposing data limits in the US | Ars Technica). And let’s just assume, with their monthly allowance and monthly fee included it all averages out to be just $0.1 per GB(let’s say)

So for it to all make economical sense lets convert it all to GBs. lets just say one hour of video is 1 GB for simplicity. So, you have 18k GB of data wireless data streamed every minute, and then 80k GB of data also streaming every minute, costing the farmers about $9800 every minute, just for bandwidth. You might argue that the 80KGB of netflix streamed overlaps the 18kGB of wireless data, but I did not even calculate the WIRED data as well as the amount of YouTube videos watched every minute(which is about 20k hours actually, so just know that i definately did not over assume.

Then, you have new people uploading data, at about 200GB per minute. So that’s a cost of $49 per GB uploaded just for farmers to break even. I don’t know who in the right mind will pay that to upload a 1GB video to a site like YouTube, so newly uploaded data to the network will be even less, say 100GB, then it’ll cost $98 per GB to upload new data to the Safenetwork. And if it goes like this, it’s a positive feedback loop which may severely damage the network, it is this i was afraid of. So instead of making it a one time payment for all data, make recurring payments possible as they’re needed.

And also make it available that website creators can pay for it’s users. Because let’s say you have social media, do you honestly expect people to PAY everytime they post something? If not, a social media site is impossible on the safenetwork due to the economic model. But if website owners are able to pay for users and make it’s users freely upload stuff, they could make money from the advertisements just like Facebook does.

Another thing to do to make social media possible and the network better in general is introduce the concept of Gigabyte-Hours, so one gigabyte-hour can be a “credit-token” given to someone who have donated 1GB of their space to the network for one hour. And this credit then allows them, to, and for free, store 1GB of their data on the network for one hour.(it could be different units like megabyte-days etc etc) But the concept remains. This would also be good for the network and make it more sustainable. Not saying it won’t sustain now, as there probably will always be people running vaults even if they lose money, but if the positive feedback loops gets going, there will be a point where the existing farmers are contributing WAY more bandwidth costing them way more than their income that, even if they love the network, they’d have no choice but to close down before their ISP takes their house for them unable to pay the massive bills!

1 Like

I hope I am not making you react here as opposed to thinking deeply (as I believe you have been), if so I don’t mean to, just searching for the next issue to resolve if it exists (they always do). Anyway it’s a discussion not a debate, I hope you think so to.

You seem to be shifting much more to bandwidth than storage, which is fine, but do you not think that with no new data the number of consumption only users would be dramatically down? In the image there even smapchat makes all the story of “only new data” as it is supposed to vanish, facebook with no comments would fail. Google with no searchers giving more data to them (even unwittingly) would lose advertising revenue (its business), twitter would not work well with no new data, Amazon needs data creation to sell (make and order === create data) and more. I cannot see any business there that does not create data.

Regardless of that we need to protect old data, so we agree there. The cost of delivering that data is what is paid in SAFE, regardless of old or new data, but yes it does depend on creators of data continually being there. If they were no creators then we could make all data immutable and have several huge data storage providers give access to copies of the same data in competition with each other, to prevent corruptions of data or denial of access etc. You could perhaps even fit all data on one crystal or something soon and just read it.

If this were the case safecoin creation would have to be on Get as opposed to Put which is not a big step, but probably an unlikely one. A healthier mix may be work looking at as well. We do debate that in house. Good to keep considering though, I agree.

9 Likes

I pay nothing for uploading.

Obviously farmers have to decide for themselves if they run a vault if the uploading costs them too much.

There are plenty of people to farm that do not pay for uploads (or downloads)

So you are using an extreme cost example that has the counter of zero cost.

EDIT: the following is a mistake of mobile and landline prices given by @foreverjoyful
Only using your extreme example of mobile data costs. Most farmers will not be mobile users and only when on wifi will mobile phones ever farm.
Only using farmers that have to pay for their bandwidth and only when paying for extra bandwidth. Most farmers will likely be below their 3TB anyhow so pay no extra, or are on unlimited. Or like in Australia are rate-limited to 512K/512K which means their vaults remain off till next month.

So your analysis doesn’t add up because of your use of costing that will not be used often.

Millions of Australians have unlimited cap on their internet plans and/or have access to going on such a plan. How can you use the most extreme costs to justify something when most farmers will be on umlinited cap or very large caps that won’t be used up.

2 Likes

Yeah yeah, totally not, i was bringing up those questions for you to think as well, i don’t know the answer to them! I guess it’s a bit hard as you can’t hear my tone of voice over the forum. i actually just researched myself after typing that and came up with the links/calculations

that’s absolutely true. Altho some data, if not most, are not created by them but their users, and currently their users get to create and have them there for free. For example SnapChat, If they have to pay every time they snapchat a story, i doubt their market capitalisation would ever dream of hitting $28 billion. But anyway! The point is tho, just thinking philosophically, websites are possible because website owners make money, or at least don’t lose money. Otherwise they’d have no money to pay for domain and hosting etc.

Right now as the economical model goes, website owners can make money still, but that money does not go back to maintaining the website(aka paying back the farmers). For example, you may very well have a YouTube on the safenetwork that makes a lot of money with advertising, but the owner really, just has to design it, and pay a one time fee(then maybe do some marketing) and off he go sleeping for years, the people who UPLOAD the videos will pay money, the advertising will bring him money, but he doesn’t need to pay any money if he doesn’t want to to the farmers. The whole thing seems unlikely it’ll work out. UNLESS you make one of those following suggestions i suggested(such as make website owners able to pay on behalf of it’s users, or introduce Gigaybyte-hours credit token) , or find other solutions, such as maybe convincing most ISP to give unlimited data(probably a much harder job, but some ISP do do it i think, although i don’t know if they have a fair go policy and shut you down if you abuse it)

3 Likes

OHHH, is that website for MOBILE data? I thought it was for landline, when i was reading 3TB data cap per month then $10/50GB I thought no way it can be mobile data… But if it’s mobile data then please disregard my calculations. I’ll have breakfast then come back and calculate again using the landline prices which I’ll try my best to find.

3 Likes

Certainly not a big step since this what its mean to be now. Isn’t it? :wink:

1 Like

Yea good point, I was confusingly saying that gets create safecoin payments, but puts supply the safecoin to pay out. I should state it more clearly :smiley:

1 Like

Sorry not necessarily normal mobile either. Sorry, so its not the most expensive. 3TB caps would be landline.

But my point stands that any amount being charged for bandwidth is infinitely more than what most Australians can have access to now.

So no matter what figures you use its way more than what is available and its up to the farmers to decide if they farm or not. So the majority of farmers will be on the cheapest bandwidth available, mostly free. Europe is similar too if I read some of the posts correctly with 1GB links up/down unlimited

1 Like

This is true, but I bet the vast vast majority make money from the info they provide customers who then use their service. I mean plumbers, joiners etc. Then activist sites are paid by activists to promote a story etc. So some sites are the business (Google, FaceBook etc.) and these are different beasts, a new business model that kills some older models, like the bookshop, who would have had a site that they paid for and made no direct profits form it, but indirectly hoped to get customers through their doors, which they probably did, but nowhere close to Amazons supply chain and fulfilment capabilities on huge scales.

tl;dr Most sites cost money, not make it, but indirectly fulfil a purpose which may be profit, indirectly though.

More dynamic sites, can be considered almost apps, with customer created content and actions, occasionally interacting. Now then you have a whole ball game we cannot value as it will depend on the app devs ideas about sustainability or profits.

I feel this is true, but if you don’t then we could be at an impasse we need to resolve there.

1 Like

But David, so it would make social media as we know it impossible on the Safenetwork, or at least people need to PAY to post their status on social media. Why don’t we change it a bit so that the current social media system can be maintained as is on the Safenetwork?

I think majority of people who are not doing illegal stuff will prefer the current Facebook, instagram, snapchat over similar ones on the safenetwork due to the fact that they’re free, even with the privacy concerns - which most people don’t really care about.

True or not, I am not sure. But if you have proposals that’s weakly dominating(that they can either make the network better off and at least no worse off), why don’t you consider changing the network a bit instead of wanting to preserve the current state? it’s still at its early stages of development isn’t it? it would take much harder effort to change later than now .

These already are in demo apps and projects on SAFE. “Impossible” is another of my never use words :wink: It is possibly to do this in a myriad of ways. Today somebody pays, usually the app supplier, that can still happen. In SAFE there are a lot of ways to also do this, i.e. read X things, get a credit (that pays the safecoin part) etc. App dev then pays, but the user increases the reach of the site and then possibly again when they post content. So monetise that directly and not in a hidden sell the user’s data type approach.

The issue/trick is using new tools to do things in new ways, not new tools to mimic old ways. The hope is some of these new ways are actually better user experiences than the current models. Privacy aside (much larger convo).

Of course another significant issue in SAFE is that publishing a site/app is actually global. So a dev in country X creates and app, now people in country Z do not worry that country X located servers/companies are sniffing data. Users may not care in some areas, but business certainly does. So cross border data handling becomes a solved issue. That is also a huge driver and B2B is an enormous market.

Probably straying off topic now though.

4 Likes

Haha fair enough! But i did say social media as we know it today, if the app developer pays for it it’s a little different than how it works now. And yeah i guess lets just set privacy aside just this discussion, but either way, i feel like the cost of maintaining a certain piece of data should be dynamically changing based on the amount of new data that is coming in. Because the idea of paying ONE time for perpetually being able to access that data, and get as many downloads and bandwidth as possible, is hardly-possible. Its much like perpetual motion, you can’t push a wheel once and expect it to run forever, that’s the underlying logic behind it, hence my concern. Something has to be giving it extra energy, at least that’s the universe as we know it today and i feel like it’s similar in the ‘economical universe’ so to speak.

Also something to keep in mind is as technology advances, the data storage cost will also be cheaper, aka cloud storage services will be cheaper and safenetwork has to match or at least be correlated to them in order to be successful, so the decreasing cost in bandwidth argument is also matched with the decreasing cost for alternative data storage.

Right now it’s purely driven by people’s desire to store new data on the safenetwork and paying for it, that’s what’s making the ‘wheels’ turn, but I am suggesting it should also be helped by those who have a desire to maintain their data on the network, that way, it can keep turning even when the people who are turning the wheels to store new data on the network gets tired, if that analogy makes sense. Because you can’t guarantee there always be the same demand to store new data as the demand to access existing data, as more and more data gets stored on the network, THINK IT THROUGH THOROUGHLY THOUGH - there will be more and more demand to access the existing data, and there will come a tipping point where the demand to access existing data(the bandwidth cost) is more than the demand for storing new data on the network(the amount of safecoins that get’s paid to the farmers) So then what happens? Some farmers might pull out, then? there’s less space on the network, and storage becomes more expensive, resulting in even less demand to store data on the network, then more farmers pull out, resulting in even more bandwidth demand from existing farmers as well as higher cost to store new data. This positive feedback loop is extremely dangerous.

Right now people are arguing that tipping point will never be reached, and that there will always be not only new data(which i agree), but those new data will also be paid-for for storage onto the safenetwork specifically. Sure, this may happen, and it’s fucking awesome if it does, and continues to happen forever. BUT, i’m only suggesting implementing some type of safety net, as to prevent the positive feedback loop i described which will lead to harms to the network and everyone’s precious data on it, if this doesn’t happen as we expected!

I realised i missed something on this, and that is, there should at least assumed to be a so called ’ base demand’ for storing new data on the safenetwork(at least i hope) so let’s say there will ALWAYS be people willing to pay basically any amount to store NEW data of lets say around 1TB per day onto the safenetwork across the world, they will support the network such that if farmers are losing money and pull out and storage becomes expensive, it’s incentive for some farmers to go back in. BUT, this comes back to my other(similar) argument - is that having an economical model like this will LIMIT the growth of the network to the demand of storing NEW data on the Safenetwork. The growth of the network becomes a ‘slave’ so to speak to the demand of storing new data on the network. This can be prevented easily by making people pay on a recurring basis if they have not stored a certain % of new data per month in relation to the existing data they’ve stored. This will allow the network to expand with much greater limits!

1 Like

I appreciate the effort you put in to define the issue in more detail.

From another angle, spinning also on one of my previous posts and what @neo said about data:

One of the deep implications of current model is that you actually own your data. And this might be a small thing to say, but it is a huge thing if you start thinking about it. How much in this world can you really feel secure about really owning?

It’s not much. Everything can be taken away from you quite easily, and it repeatedly happens whenever there’s a slight turmoil (many examples and different situations).

Holdings on bank accounts are just debt notes. Property, well as long as you can defend it, (be that delegated with paid taxes and a sound and working legal system). But these things erode in times of trouble.
Every physical thing you own, is yours only as long as you or your delegates can defend it.

I think with SAFE network, the network itself, as long as it lives, guarantees that only you own your data, no matter what.
That, is IMO huge.

In the end, we have same goals though. You seem to have the robustness of the network in mind. Throwing out one of its fundamental and most revolutionary properties, would be something we want to avoid. Else the concern of keeping the network alive is suddenly not that large anymore. IMO.

EDIT: For clarity I might add, that the rental system you propose, implicates that you do not own your data anymore. It can be taken away from you.

7 Likes

This is very important. Also the potential for loss of “libraries” of valuable data simply because something happens to the maintainer. (death, loss of job, illness, got bored, etc) And the 404 errors because as we know today not paying hosting fees, or people moving onto new things we see the internet littered with 404 errors. I find this a lot when trying to find datasheets on chips etc. Forever data will at least mean these old pages (maybe just 1 year old) and now lost pages will be greatly reduced and only reason for a missing page is the safesite owner deliberately removes the page from view.

Also another point is that if you have to pay rent on data then that data has to be linked to you, For immutable files this means a radical change to the security/anonymous model SAFE works under, The immutable files are not linked to anybody at all and to apply rent then they have to have a link to you and thus removing a level of anonymity.

EDIT:
Also for immutable data this means comments on forums or blogs will disappear regularly when people don’t pay the rental on their MD objects used for forum comments and think of the facebook replacement, it will become a dearth of missing post/comments/images/etc What about the SAFEcoin I have stashed away, it is also now a piece of data I have to pay rent on and will it disappear if I forget to pay rent on that object.

So for the potential 100’s of thousands of MD objects I have not to mention the millions of immutable chunks, I have to pay rent on. How is anyone expected to keep track of it all? That in itself is a project of no small size. Network modules to be created, network modules to be changes, APPs to keep track of when rent is due on which object. Look just charge me double the going rate to remove any rent.

3 Likes

You still ‘own your data’, it’s taken away by a ‘smart-contract’ so to speak. You have absolute say whether it’s taken away or not and know the exact conditions that needs to be met for it to stay or for it to be taken away.

With the current model, sure, you own your data and no matter what it can’t be taken away… unless of course, the whole network goes down or farmers quitting leaving your data potentially extremely vulnerable to disappearing forever if there’s not enough space to keep the 8 copies, i mean sure, even at 4 copies it’s still OK but it’ll just depend on probability. So you own your data… based on a probability function… you may not ever see it again, but if you do, you own it… still a bit unsecure i think…