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

The SAFE NETWORK can be updated by core developers as needed. If issues arise on any level devs and the greater community will mull over potential solutions and implement them when agreed upon.

Please create a formal proposal that demonstrates a thorough understanding of the network protocol, design choices, and the impact (bandwidth overhead, code complexity, security, anonymity, etc) your proposal both negative and positive will have. We’ll go from there.

ATM you’re being orientated as you debate. Its exhausting for both spectators and participants of this thread. Be fair and do your due diligence. Return when you’re ready. Please don’t troll as would be the case if you continue prematurely.

3 Likes

Sure, although I’m mostly just bringing up this issue for discussion. My proposals are only basic ideas, and is up for improvement and discussion by the community.

Oh really? I am wondering, Once the real network has launched, can it still be updated without taking down the whole network first? I would like to know. Because if it can be updated as it goes without compromising everyone’s data, then I have no problem with the current model proceeding as is. As you can change it later anyway. I am only bringing it up now as I was afraid this won’t be the case.

1 Like

Because I was not addressing that but the issue of your charging people to simply browse. You talk of others debating better, then you had better also rise to the occasion and not use cheap tricks to dismiss someone trying to educate you about the network.

Once they are updating *anything* then they have a SAFE account and have at least bought one coin’s worth of resource. With that one coin they can update facesuck pretty much as many times they want for the next decade or two.

They are already paying their internet provider to access the internet and if they want to upload (or update facesucky) then they know it will cost.

So browsing being free covers the billion or more of people who never upload anything, never write to social media or write to forums. Which is BTW one of the stated goals of the network to provide FREE information to the world. As opposed to the millions (maybe billion) who “write” to some part of the internet.

Again though for those who do “write” to the internet how many are just/mostly doing their banking, the odd update to social media, and emailing??? I suggest that most are only doing that and the one coin to set up their account will last them for years to decades.

So its on a different level to having to set up an account just to browse and the overheads for the code and farmers with this will be higher than the simple pay once forever store which is the cheapest method.

That doesn’t even hold true for too much else. For everything it depends on what the second option is and the effects of it.

Town water. Oh lets allow a second government department (your rental system) to supply town water. This will allow them to charge by a different model and to do so they have to install infrastructure (More code and then procesing) and process the water liter by liter (check each chunk for deletion each 1/30th period) and then lay separate pipe to each house to give them the option to choose.

Now this outlandish second option is not better is it. Everyone would end up paying more in taxes and water charges. So your rental has some parallels with this in that extra code, extra processing that every farmer has to pay to run because of the higher cpu usage and disk usage. Its that miniscule amount increasing and over the whole world results in higher costs and farmer rewards would have to be just that bit higher.

Then the cost of lost data because some (many? most?) people WILL forget at some time to pay the rental on some of their data. The network is no longer seen as secure and reliable, usage will drop off. Then after they have uploaded their photos a number of times then they will be spending a lot of time just sitting there clicking to pay the rentals at the right time. Since the photo sets were uploaded at different times in the month this sitting could be 5 or 10 times for a few hours each time for every month.

Well they are going to be paying the right amount to upload from the start. THAT is what the upload cost amount/rate is ALL ABOUT. Yes I suggested it might be an idea when I read his proposal, but in reality the upload cost is all about charging enough to meet the download costs (on average over decades). While it is not a direct upload payments pay farmers, it does reduce scarcity which increases farmers rewards - in other words its a complex linkage and not a pay to upload to pay farmers. More like put your cup of water in the lake and the farmers use their cup to get water when allowed - but yes more complex and not direct linkage/parallel

I think you still don’t understand this at all. REALLY I don’t think you understand the model. You compare apples and oranges then come up with a conclusion and claim others don’t look at this from another perspective.

Its like you look at something through dark glasses and then say everyone else is not seeing it right because they use clear glasses.

Its a new model, it does not follow the dying economic model in use today. Google is sucking money out of *you* because they shape *you* through advertising and channel *you* into certain sellers, certain economic models, way of thinking. All of this makes people buy more from more expensive people and they skim off the top through the charges for advertising.

There has been considerations and as I told @foreverjoyful the discussions are somewhere in this forum. One aspect is the increased payments to farmers when resources start to drop off. And if it gets worse the payments grow and grow fast as it gets more necessary. Human nature kicks in there. But obviously if it got as bad as losing data then rental or paying to browse is not going to help much and is likely to contribute to the speed of the demise because people give up quicker the more they have to pay.

Yes of course. The network will gain future features through such a mechanism, and David even mentioned this as a priority for the upcoming alpha/test networks.

4 Likes

The debate is heading toward the safecoin distribution algorithm which is currently not designed, so both sides of the debate are currently theoretical. With that fairly large caveat aside, there seem to be a few points of misunderstanding.

I disagree. The sustainability is independant of new data.

Assume (dubiously) the worst case scenario where no new data is uploaded so no new coins will ever be made available for farming.

There will be some amount of coins available at that time. That amount of coins available will always be decreasing as new coins continue to be claimed via farming. But the amount available never increases because nobody is recycling the ones they have. This is something I think we both agree.

Coins from the pool of remaining coins is issued to farmers at a variable rate which adjusts intermittently, very much like the bitcoin mining difficulty rate. This is speculative, since the algorithm isn’t fully designed yet.

The wiki is old, but since the feature hasn’t been designed it’s still the most relevant info currently available. [Here’s the current idea]:

https://safenetwork.wiki/en/Safecoins_(How_it_works)

The earning rate also takes into account the rank of the Vault, a process whereby the network scores the usefulness of each node from 0 (being the worst) to 1 (the best). The safecoin farming rate is ultimately the result of the network rate, a balance of the demand and supply on the network, multiplied by the vault rank.

The network automatically increases farming rewards as space is required and reduces them as space becomes abundant.

So there is an adjustment mechanism for the rate the remaining coins are issued to farmers. This means that no matter how much the pool of available coins shrinks, there will always be some coins to farm for (even if they are almost impossible to obtain). Thus there should always be some incentive to farm (beside the obvious incentive that farming keeps your own data safe regardless of whether you’re economically rewarded or not).

This does pose a problem though: there’s no incentive for users to recycle coins (which increases the number available for farming). I think is this is the basis for your perspective that farmers depend on users spending their coins for farming to be sustainable. However, this is not true. Farmers only depend on users spending coins for the available supply of coins to increase, but that alone does not necessarily make farming unsustainable. Farming can be sustainable even if the supply of coins never increases. I would be keen to see proof of that being incorrect.

To return to a bitcoin analogy again: Newcomers to bitcoin argue the mining reward eventually being zero will kill the network since miners have no incentive to continue securing the network. The standard retort is transaction fees will sustain the network. The argument being proposed in this topic for the safe network is akin to saying ‘what if nobody transacts on the bitcoin network once it depends on transaction fees’. There’s a bigger problem than no transaction fees - ie there are no transactions! Not to say this isn’t a problem, but is it real? We can only speculate, never really know for sure, despite our best intentions.

See above for the safe network design that also has an adjustment mechanism (albeit currently not a particularly clear one).

I try to debate the ideas not the person, but the lack of understanding of this proposed mechanism really does make it hard. You need to read more about the network. The wiki article is a pretty obvious thing to have read when discussing safecoin mechanics at such length as this, so you can see why a lack of basic background work causes doubt of your intentions. If you really cared, you’d have read more.

Not true. The difficulty remains high until the next retarget. If 99% of miners have stopped mining then blocks will be extremely rare and retarget will take a very long time. What effect this has on users and the network, well, we can guess pretty well but until it happens we can’t say for sure. Just like with this topic.

This means the network would not be autonomous and decentralized so I doubt this path will be taken. But for sure, it’s possible that human intervention may happen at some point, I just think it’s against the ideals of the network for this to happen. (Is that a political shitstorm I see approaching?!)

6 Likes

There is currently the RFC that I have relied on and the stated goals by David and some others. These I have used for my basis.

Yes I agree with this. Although there are effects that flow though

Actually wouldn’t the current RFC be a better source.

The White paper suggests that this payment of farmers could continue for decades with the reward success reducing as the available coin becomes more scarce.

Of course as we see elsewhere that the more scarce the available coin to farm (mine) then the more valuable the coin/item becomes assuming people want it in the first place.

So as you said the more farming is done and the reward success reduces then those few coins farmed in a decade or two are worth a lot more then the many coins farmed today

I disagree autonomy does not rule out the owners turning off their part of the network. Autonomy does not rule out control of this sort or manual intervention.

You could say we people (as a “machine”) are autonomous. A person could grow their own food and look after themselves and decide for themselves within the bounds of their environment how to live. Even in a city we (as a “mechanism”) have what can be called autonomy and rely on others to provide things we consume.

The network relies on people to provide the machines, keep the electricity connected and the links supplied. And still be considered an autonomous network.

Then like people sometimes need intervention (surgery) so too the network will require such intervention. But we still consider that autonomy.

The process of upgrade initially will be people stopping their nodes & upgrading while the network still runs autonomously and if enough people do this then the network will be upgraded. David has suggested that eventually the network will be able to test for itself the new upgrade and if found suitable the network nodes will upgrade themselves. Will that happen, well it will if David can work out how to do it.

3 Likes

Indeed, good point.

0004 - Farm Attempt as posted already by me in post #8 of this very same thread

This result [of the farm attempt algorithm] is then modulo divided by the farming rate (FR).

0012 - Safecoin Implementation also in post #8

the farming rate is defined as

I admit these documents are difficult for non-technical users, but that means more reading is needed, not less.

1 Like

Wait wait wait, how is this apples to oranges? So right now YouTube makes money by ads and they pay for their servers and for giving free bandwidth to people to view their videos. As well as server security, data security/backup, server maintenance and DDOS protection etc etc etc. Now, with the safenetwork, all of those issues do not have to be paid for, and they can still benefit from advertising if so they want to. But those bandwidth costs will still be paid by farmers, as now if they want to view a video a billion times, they do not strain YouTube server’s bandwidth, they use the farmers bandwidth, which cost money, and in return, if there’s an ad in front of the videos, YouTube, using the safenetwork, will still be paid money, while not having any ongoing server costs, just one time upload fee, but they will get paid on an ongoing basis because people are using the farmers bandwidth and viewing the video.

From what i describe above, you don’t think that’s draining money out of the farmers?

If it charges enough for decades, then it’s probably a bit too much than what people are willing to pay for, hence a optional rental model will definitely be a more attractive option anyway.

Ok, but WHERE do the payments come from?

That’s the most soothing thing i’ve heard so far. I made a wrong assumption at the start and that is i thought once the network launched, it has to work(that’s why they’re taking a long time), this is what i have heard from the host of a meet up I have been to, hence why i got this assumption that the network once launched, it needs to work and can’t be changed. if it can be changed i guess I’m not in a huge hurry to propose a change. We can see how it goes first. Nevertheless i do think adding a secondary option of a rental model will not damage the existing model in any way, it will only benefit it. (although TECHNICALLY i am not sure, if it need much more resources/bandwidth to keep in track of the data expiry, i am only talking about it in the broad economical model sense, it doesn’t do any damage so far from what i can think of)

1 Like

this is how

If you don’t understand why, then read all that has been said to you above and in the other thread.

Well yes from what you wrote I guess it might. But that doesn’t mean what you wrote is correct.

It uses the *forever* model in that one respect and we see this in other offerings by companies.

Burial plots
Lifetime memberships
and so on I will not waste time here, its all been said now.

The algorithm would need to pay more and make rewards more scarce more quickly, thus speeding up the demise.

In another post by @neo

Ok first, I just want to clarify, if you say the sustainability is independent of amount of NEW data, that is to say, even if there’s no new data FOREVER, the network will still be able to sustain it’s existing data as those factors are independent. I think even @dirvine would disagree with this, from how he explained new data gets coins returned to the network to pay the farmers.

And Neo also disagreed with this, he said even if there’s almost no new data, then the network dies. I disagree with him in that the network could still very well be used for people accessing their data, but farmers will slowly quit and hence die. So there’s a seeming contradiction between what you have said @neo, could you clarify?

This is also the same as saying - if Bitcoin has reached it’s coin limit of 21 million, and the only income is from transactions fees, and if say the total amount of transactions fees total halves every month as new transactions decrease, there will be no effect to the Bitcoin network. This is outright impossible. The miners pay a certain cost of electricity, if the supply they get keeps on halving, one of these 3 things must happen, or a combination of the 3, as a result of miners getting paid less Bitcoins.

  1. Miners finds more energy efficient machines that cost less to mine the Bitcoins. Some can find free electricity from solar panels and/or invent new ASICS or quantum computers that are hyper efficient.
  2. Bitcoin increases in value and to miners it does not matter if they get paid less Bitcoins as their value is the same or more.
  3. Some miners will quit mining and hence makes the Bitcoin network less secure as they no longer make a profit and make a loss(or they can choose to continue at a loss for philosophical or other reasons, but not many people will do that)

Contrast this to the safenetwork, if close to all 4.2 Billion safecoins are mined and only few remaining, no new coins gets returned to the network as and halve of the remaining few gets farmed monthly, directly contrasting to my previous 3 points, we have(and one or a combination of these things MUST happen)

  1. Farmers will have to find more efficient ways, like you have mentioned, this will definitely happen for some farmers, bandwidth can cost less, electricity can cost less which means storage will cost less, some can get free electricity from solar panels and/and have big farmers getting enterprise ISP deals for very cheap bandwidth.
  2. Safecoin increases in value and to farmers it does not matter if they get paid less Safecoins as their value is the same or more.
  3. Some farmers will quit farming and hence makes the Safenetwork less redundant and potentially lose a lot of data, as they no longer make a profit and make a loss(or they can choose to continue at a loss for philosophical or other reasons, but not many people will do that)

So I’d argue, with the current economical model of the safenetwork, it is definitely DEPENDENT on NEW data being stored. In fact, my proposals are aimed at making the safenetwork independent or MUCH LESS dependent on the need for NEW data to sustain the network. If you going to say it is independent you really have to have extremely good arguments backing it up, as from how @dirvine described to me about how coins gets returned to the network only by people storing (new) data on it, it’s very much so that it is definitely dependent on new data being stored on the network.

Isn’t it also (at least partly) the reason why @neo proposed the Pay the Producer model to incentivise people to upload new data to the network constantly?

1 Like

Considering this is not what @mav or myself were referring to then this is a NEW question and requires a different answer to what you quoted.

What we were referring to is if farmers continued to farm then they will be paid and if SAFEcoin is worth something then that farming would be profitable. SAFEcoin if its worth something to people then as the farming rewards become more scarce then the price of SAFEcoin would increase and thus farming remains profitable.

Now the question of if farmers would continue then that is a complex one and many scenarios are possible. Some where it dies and some where it survives for a decade or two and beyond.

Personally I would be looking at the “people will support the network they use” scenario This is what we see in life. Community groups, clubs, charities, etc where people support that which they prefer/use. So if people are using SAFE by being farmers then they (in general) will also be using SAFE for their storage, social etc needs and thus uploading data. And for those who are using SAFE to store their data, social media etc will if they can be farming too.

So the scenario that SAFE is not being used to store data and farmers continue to farm will be contrary to observed human nature and thus is unlikely to occur. Either in general people will stop using SAFE (store & farm) or be using SAFE (store & farm) <---- not necessarily each and every person but in general sense or on average.

No its not because they are different beasts with different outcomes because of the difference between the two. Drawing a parallel this way doesn’t work.

This is like many decades for this situation to occur. So makes its impossibe to do this sort of comparison, the time scales is an additional factor in the comparison. It affects each differently

Again this is not what we were addressing but yet another issue. Your statement is a global view and still is wrong. The network can survive for decades with no new data if people are still using it significantly which means farmers are still paid for decades and the coin will follow the path of increasing in value due to the difficulty of getting any new coin. People would have to be hoarding it (not spending it back to the network) and changing hands at times. But this causes scarcity and price rises thus the smaller amount of coin farmers get in 2 decades time still is very profitable.

No that isn’t the direct reason and not the real direct purpose. The major reason is to get people to use the network. Oh it was David that proposed the PtP concept. When people use the network then they use (store and farm if possible/desire to) the network.

That’s the definition of being INDEPENDENT though, @neo… i don’t know how it is a new question. By saying it is independent, the case i described HAS to hold true. Otherwise the network’s sustainability is DEPENDANT on the amount of new data being stored. True @mav?

Because the process is not the traditional model and lack of understanding makes it seems that way but when you understand it you see that the question is actually different in scope and result.

There will always be new data. Imagine IoT connected to SAFE, or Tesla cars because the cars create gigabytes every few minutes, this isn’t to mention what if this conversation was on SAFE? We would have no shortage of new data because this thread has gone in a seemingly infinite loop. I’m only being cheeky (kind of). Anywho, not that I know much on the subject but this idea is borderline ludicrous. It’s a complete hypothetical that is basically infeasible. If there was no new data it’s because there is no one to put it there (human extinction) and therefore no need for the network. I just don’t see it happening. Here’s another hypothetical. Say the whole world uses SAFE to store their data and no one ever creates anymore data but everyone wants access to old data? Storage, and computation will be so cheap and abundant that SAFE will just run in the background still securing and serving up data unnoticed on the worlds devices. This goes without even bringing up the fact that the network can be upgraded and if there was an existential risk to the network we could solve it. Heck the team hasn’t even gotten to run simulations yet, not that this discussion shouldn’t happen but dear lord the splitting of hairs is unnerving.

2 Likes

Exactly and part of what I was saying about if people are using SAFE then they are going to use SAFE.

1 Like

I think you have to zoom out here.

If someone uploads a popular video and lots if people view it, who his benefited? Largely, those who have enjoyed consuming it. Who have profited? The farmers who provided the data as they were awarded safecoin. Who is unhappy/cheated? Nobody, as everyone has had a positive trade.

Yes, there may money made through advertisement. Maybe some of this money goes to the person uploading the video. Again, who is winning and losing? Portal site gets money for providing a service and the producer gets some money for their efforts.

Proportionately, costs and profits will adjust. If the portal overheads drop, so will the barriers to entry. Maybe that means cheaper videos all around or maybe more competition, but either way, people are happy.

I think people get lost somewhere around data ownership of data. There seems to be a belief that the owner must pay proportionately for the data being accessed. Ask yourself why? If many people are getting value from reading the data, then why should only the uploadet pay? Moreover, why should farmers care which data is popular and who uploaded it, when they have no notion of what data they even served.

The bottom line is, the network should deliver data which people value. It doesn’t need to pick or choose why or base it on who uploaded it. The network is agnostic, it is neutral. The cost to store something just goes in the big pot and then pays out for the things people value most. This is a beautiful thing.

1 Like

I’m tempted to post this to a new thread as this one will soon be long enough to reach Mars, but since some knowledgeable heads are around I’ll leave it here for now.

Can someone tell me if this description of the mechanics of farming, leaving aside the economics for a moment, is broadly correct?

  1. A GET (a request for a particular chunk of data) is posted to the network.
  2. All of the nodes (vaults) storing that chunk race to fulfil that order.
  3. The node that delivers the chunk quickest is rewarded with Safecoin, the rest get nothing.
  4. The factors that influence which node is likely to succeed are (in order):
    a. the node’s closeness to the requesting client (in XOR space);
    b. the number of copies of the chunk that are cached;
    c. the node’s closeness to the requesting client in physical space (because of latency);
    d. the node’s record of reliability and trustworthiness (node age);
  5. The payout to the successful node will depend on a variable called Farming Rate which is dynamically adjusted to maintain the optimum level of free space on the network.

If this is correct, I am struggling to get my head around part 5. Does the value of Safecoin fluctuate according to the Farming Rate variable so that each successful delivery of a chunk is rewarded with one Safecoin no matter what, but that coin’s value is variable. Or does Safecoin’s value remain broadly stable, with Farming rewarded in fractions of Safecoin (ie fine-grained divisibility) so that one day delivery on a GET request might be worth 1 Safecoin, the next day 1.1 Safecoin? Or is it the case that not every successful delivery of a data chunk following a GET is rewarded, just a certain proportion of them as determined by the Farming Rate?

1 Like

Yes, i agree to that, there wouldn’t be a shortage most likely, but non the less the sustainability of the network is DEPENDENT on that fact. Which is what i’m trying to change a bit, since I feel like the network shouldn’t have it set up in a way such that it is DEPENDENT on new data. I’d very much want the network to sustain even if no new data is being created or new data significantly reduces, cos who knows what will happen in the future. But in saying that, because how @neo has said the network can be changed anyway once it launches without the need to compromise existing data on the network, if that’s the case, I would be much more confident in everything and will definitely recommend this to more people and buy more Maidsafecoins myself :slight_smile:

AND the website who those people have went to to watch the hosted the video, due to advertising ETC, and the farmers get nothing out of the advertising profit by providing the network and making everything possible in the first place.

Well, because of sustainability as i’ve mentioned. Who WILL pay for the bandwidth costs then?

It is absolutely fabulous, don’t get me wrong, i’m simply bringing up some sustainability concerns for discussion. Although I’m not too worried now as I’m told that the network can be changed even after launch without compromising anything(Am i right in saying that @dirvine?)

if

1 Like

Like i said there may well be new data forever, but that doesn’t support your argument at all of how the sustainability is INDEPENDENT of new data being stored.