A couple of useful places to start:
To be honest I’m not sure about this point “Users can maintain websites at low cost, regardless of how many hits the website receives”. If some content is having a lot of requests, it has a cost for the network (the Vaults are being paid for that) so who is paying for this? The inflation process takes care of it alone?
Thank you, I’ve read it and trying to summarize.
The network is paying, this is what farming is in a nut shell and the whole point of it.
Beyond that the site should get PtD (Pay the Developer) rewards in the form of Safecoin when the site is visited. In what amounts is up in the air AFAIK.
Publishing a site is extremely easy
The publisher doesn’t have to worry about paying domain hosting fees, security, or hosting their own server
Opportunistic caching will also keep data closer to where it is being requested so as a site or app becomes more popular in those moments when a centralized server would become overloaded, the data will accumulate and be more highly available and basically have the exact opposite effect where the site or apps popular data gets to the destination even faster.
With perpetual data that effectively prevents link rot and acts as the internet archive where you can go through previous versions and edits to any Public data that has been appended to.
This is just some simple unique things off the top of my head. There is much more so glad you are digging in. Check out the safe network primer linked at the end of the dev updates as well.
Nice, you are helping me a lot. But there’s still a gap in my understanding. Let’s say Bob upload a file to the network and paid 5 SAFEcoins for it. Then this file became very popular so a lot of people are requesting. In my point of view, that requests are requiring a lot of work for the Vaults/Farmers (they need to be paid for that). So in this scenario we could have 5 SAFEcoins being paid and 500 SAFEcoins being demanded. Where is the trick? I’m not visualizing this “opportunistic caching” you mentioned because we will always need the work of the Vaults. Sorry my lack of knowledge.
I’d say data interoperability between apps will be a big one. Instead of your data being locked in a proprietary silo and you being at the mercy of whatever curation algorithms are imposed by the silo owner, your data is yours and how you interact with it and link it together with other users data will be under your control. If you don’t like how the data is being curated you can easily switch to a different curator. Everything is very mash-up-able (is there a word for that? Just interoperable I guess?)
I did a bit of a proof of concept back in the day
Let’s try to run through the resource based economy.
Vaults (nodes) make up the network and have many personas. They store data chunks, serve data chunks, their age rank increases with expected/good behavior over time, being promoted from child->adult->elder. The responsibilities of the vaults personas change based on age. As a child you are not trusted with much anything by the network besides being a storage device, as an adult you can handle and store data, as an elder you will sign data and transactions and so on. Of course the software does all of this autonomously. If you constantly lose connection, fail to serve data, or do malicious hacking that is outside of expected behavior, your age will be halved/demoted and eventually can be removed as a vault. When data is requested and successfully delivered from your vault you are “farming” Safecoin.
Safecoin amount rewarded to vaults is adjusted based off from network needs. If there is not enough storage, more Safecoin will be rewarded, if too much storage is available them less will be rewarded. I believe there is some buffer of like 30% of free space across all vaults at all times. The network mints fresh Safecoin for vaults/“farmers“
When a client stores data they pay the storage rate at that given time which is constantly adjusted by the network to be as low as possible. That just means it should be as cheap as storage should reasonably be and competitive without being harmful to the network. What they pay the network to store their data, is then burned, more like buried ;D by the network. There is a max coin supply and recycling of coins helps the network and coin scale sustainably.
When a client requests any data, it is free. That is a fundamental of SAFE for anyone to have access to data for free. When a client requests data it is giving the vault/farmers part of their intended purpose in which the network rewards the vaults for doing so.
There really isn’t anything disparaging in this relationship. What you maybe are taking issue with might be addressed by the following.
In regards to opportunistic caching, as data is requested and along the route of its destination, that data can be cached temporarily by other nodes which when you think about a global network, can help make serving that same popular data you are talking about, faster, more efficient, and less intensive on the network. Vaults just have to fill their roles and I don’t believe opportunistic caching is rewarded. It’s like having any job, you have some things you may not have been hired specifically for that are required of you but if you didn’t do then you’d be fired.
Hope that helps.
Caching will handle the popular content and that bypasses farming.
At this time cache hits will not attract any specific rewards. It is part of the node functions that a farmer also performs. The farmer is paid for farming and the farming rewards should be more than enough to cover the cost of being a node incl caching. Of course for the home farmers the node functionality will not cost them since they are using spare resources.
Also some popular files will “cost” the network in farming rewards close to or even a little more than it charged for storage. But for every popular file there will be all those backups people do, all those home movies/photos that are viewed rarely, all the hopefully popular content that never became popular. All those web sites that get a few hits a week and some are large.
There is a lot of spreading the cost over all data uploaded and remember this is the network running and owning itself. The rewards come from the network paying the rewards and no one pays anything for those rewards. The network charges to use its resources (basically storing data). The store cost is calculated by the network to cover its costs and so the users uploading will be paying enough to handle the popular sites.
Now of course it is important to create self adjusting algorithms that will adequately work out the fairest upload cost and reward amounts without overcharging. Remember the network is not after making a profit for itself or any shareholders, so it can prove the storage at the cheapest cost while remaining viable.
I believe it is a matter of perspective.
If someone uploads something which many people find valuable, the network has succeeded in fulfilling the demand. It has delivered what the users want. This is a great thing.
If someone uploads something which most people do not care for, it is still available for those who appreciate it. The network has still succeeded in delivering what some people want.
While the relative cost to the network may be higher to deliver something valued frequently, this is balanced by the fees generated from users uploaded something less valued by others.
In short, valued data costs will be subsidised by the upload fees to less valued data. All uploads cost the same, but not all data has the same overhead to the network.
To get the best value out of the SAFENetwork, users should upload data which many people desire. This is a great dynamic and will incentivise the upload of data people want to consume.
We are discussing a very important topic here guys, everyone is helping me. But some points are still strange to me.
Some are arguing that chaching helps, but I sincerely believe that caching is only part of the equation. The traditional Internet already has CDNs and cache tools, but even so, there are costs for all agents.
As I understand it, the cost to upload and store files on the network will be adjusted according to the demand and work of the farmers. But I don’t know if that makes sense, unpopular files would be paying the bill generated by popular files. I agree that this is a matter of perspective, that popular websites and archives contribute to the popularization of the network as a whole, but in the long run this concept is quite questionable.
I have some websites (2 million views per month). Several times I had to change the hosting plan and pay more because of the popularity. But that was not bad, after all it is profitable for me. Popularity represents money for the creator, so I already have my reward. I can sell more products, do more advertising, all of which is a direct benefit that only the creator of the popular content has. It is only fair that I pay the hosting bill. It doesn’t seem right to me that other unpopular blogs pay my bill, after all the profit from my projects is all mine, why should I prorate the costs?
If we want big portals to migrate to the SAFE Network, surely these portals will find their ways to remunerate themselves with their audience, we don’t need to make all the people who upload files pay for the cost generated by the big portals. There seems to be a distortion.
Maybe I’m misunderstanding the dynamics of the system, please correct me if I got it wrong.
To me it’s more like make all data equal. The current model is all about current popularity. That is cool and perfect capitalist (I do not mean anything bad here) type approach to market supply/demand.
However it can miss the point of the poor persons ability to serve a huge audience. We also need to consider currently unpopular content, such as a scientific paper that can be 50 years ahead of it’s time and so on, lost scrolls etc. Same with the app rewards, we try to make it so everyone has an equal chance.
The challenge to the norm is data equality for all and pay once store forever. It challenges many norms regarding current economic models and I think it should.
Bottom line, we look to make hosting/storing data as cheep as possible and then treat all data as equal. It could be as we progress then bandwidth is still expensive and unequally distributed across the globe and we do need to consider more granular controls/costs but I am not yet convinced we need to worry just yet.
This is precisely the point. If my content is unpopular, it is cheap to store, so I pay little. But if everything is prorated, my unpopular content will cost more to me.
As long as your content isn’t popular, you don’t have to worry about costs. If you’re getting popular, you’re probably getting a chance to monetize it, so maybe you should pay more for the costs as well.
That is great!
I’m not sure if that makes sense.
currently popular, could be unpopular in future. Likewise currently unpopular could have the cure for cancer when we know more of other data etc.
So current popularity is not the complete answer to “cost”. The actual value of the data is important. So now a popular video with Gb of dance moves compared with the bitcoin whitepaper say? what has what value and is it base don bytes? This is where I feel we challenge that thinking as well as pay once being the other.
I hope I am making some sense? It’s an attempt to revalue data itself, over all time and not only current popularity.
There is also the impact of competition. It won’t just be your successful site having low hosting costs, it will be all the others too. If you are making a saving, so are they. No doubt, this will lead to an erosion of margin in time, delivering little in the way of a net benefit to you (or them).
When we think about YouTube or Facebook, we are dealing with walled gardens. They store your data and monetise it through advertising and so forth.
On the SAFENetwork, the user retains their data and can share it with whoever they wish. The apps will run locally and will collate their data and other user’s data to create a window into the data. The existing model is turned on its head.
Any sort of site which is shaping and storing data on a user’s behalf, then charging a fee or adding advertisement, is unlikely to work on the SAFENetwork. Instead, the user may pay for a tool to shape and link their data with others, but the emphasis will be on the user owning their data without the assistance of others.
I suspect their will be sites which provide indexing in some way, who could advertise for profit, etc, but the balance of power will likely shift to the user. Maybe some sites will try to tempt users to give up their data for subsidised uploads, but it will be less appealing than now, IMO.
We will only know for sure when we see how the dynamics pan out, but if the cost of entry is relatively low for all, becoming the next Google or Netflix becomes a lot more feasible - instead of needing to build out massive infrastructure, time can just be spent creating the best service and promoting it.
To add, popular non-profit sites (like Wikipedia) will be far better off. They often provide value for many people, but constantly struggle with hosting costs. They will be net beneficiaries of the network, as will their audience.
I agree that data changes it’s popularity over time. But this is why we pay different values for this data over time. The model you are proposing will charge new data for the price of the actual popularity of the old data. This old data was charged with another reference. Instead of charging each data according to it’s own popularity, you are equalizing prices and creating time shifts where some pay for the popularity of others. Why do you think this is fairer? I’m not comparing dance moves with Bitcoin paper. Both can monetize themselves or ask for a donation (depending on their goal). But if one of them generates more costs, who deserves to pay the bill? What kind of content will we encourage?
… so currently on the existing internet …
Is this accurately capturing the problem and the solution as you see it?
To clarify ‘treat all data as equal’, I feel there’s a real strength to data fungibility, all data being treated as equal. Think of it as data stewardship rather than data burden. When we see the network and vault operators as stewards of the data, then it makes sense to do things a certain way, and if we see it as a burden then it makes sense to do it a different way.
There’s an implicit assumption here that because we can find discrepancies in the SAFE model with some idea of fairness or efficiency, this system is ‘wrong’ in some arbitrary sense, but compared to what - the status quo?
It is what it is, and it has been designed based on all sorts of contributory factors, which are no more or less right and justifiable than the current systems, which are what they are. If you want to remodel SAFE to work like those, then there really is no point.
We’re here, or at least many of us are, because we want to try doing things differently and are focused on the upside of this more than we are constrained by ideas from the past.
Guys, you are trying to introduce a new paradigm about value and cost of data. It’s nice, but I’m afraid you are creating something that fights against basic market principles. There is value and there are costs. Cost is a burden. Value is great. Both exist. Each player needs to receive the correct signaling of their actions. This idea that one pays for the costs generated by another resembles socialism. “Treat all data as equal” is like “treat all humans as equal” charging the same amount no matter how much you eat.
From what I understand the data which a user wants to upload, there will be a bidding at which price farmers are willing to accept to store the data. Free float supply/demand price.