https://safenetwork.tech/fundamentals/
Let anyone browse content anonymously and free of charge
It is crucial that the new decentralised web is without barriers. One of the most important foundations for a global, collaborative platform is that anyone can access public content for free at any time without the need to create an account.
Imagine a computer on unmettered internet connection running vault, you pay electicity and internet. Serving 1 or 1000000 GETs will generate allmost zero $ cost difference for the vault owner.
From the global point of view, high number of GETs can make network slower to respond, but other than that, I dont see anything else I could call “cost”.
Just getting back to the exploding spaceship scenario… The analogy breaks down a bit because it’s not like we get one chance then need to input another billion dollars to try again. I would not be surprised if early success of the network as created by the maidsafe team triggers various forks that try different approaches. Some may embrace a GET cost for example. I mean at the very least the design IS going though many iterations right now. So the first try does not have to be perfect to avoid catastrophe.
This is a valid fear, but is it likely? The amount of new stuff pushed into the clear net daily is huge and growing. Yes, it maybe possible, but is it likely? I’d say no, it is not.
Moreover, all those dynamics at play will take time to pan out. Are we likely to see an immediate exodus of vaults if the economics turn against them? Will they abandon their hard earned reputation, which help them to access rewards? Again, I suspect not. If such things appear to be cyclic, many will tough it out and improve their status on the network.
Given some distributed systems have weak or non-existent reward systems (Such as IPFS, BitTorrent, etc), yet still survive, the model does not need to be perfect. It should strive to be better than the rest, without the complexity becoming overwhelming.
Examples in economics are present in broadband subscriptions or mobile phone packages. Do we use every minute we pay for or every MB of bandwidth? No, at least I don’t. We just pay something we think is reasonable to avail or the services provided.
Imo, the economics are important, but a working and valued network is even more so. If there is good data and services on SAFENetwork, easy ways to contribute and rewards for those who do, I’d say that may likely be enough. Longer term, things can be refined, but securing the demand for a longer term is critical.
Yes, I think the question here is: does it cost the farmer something to retrieve data chunks to service a GET, and if so, how much?
If the variable cost to service a GET is negligible regardless of how many GET requests a piece of data receives (e.g. servicing 1 GET costs as much or requires the same amount of resources as servicing 1,000,000 GETs), then this discussion is moot.
If, however, the variable cost to service a GET is not negligible, then one must determine how that cost to the system (and ultimately, farmer) would be handled.
Agreed. This is why I think it may be best to start simple (I.e. pay per PUT only), and build enough flexibility into the system to adjust if a GET fee is needed later. This is also why it’s important to understand what the governance framework for SAFE will be.
Without a governance framework, the alternative is to fork the Network and start again. This for many reasons is not a desirable solution.
While I don’t necessarily advocate starting with pay per PUT and GET fees, I do think it is worth discussing these things now to illuminate what bridges may need to be crossed in the future and how.
This happens all the time. If I’m a Bitcoin miner, maybe I’ve already bought and paid a lot of ASICs and made a fixed 1 year energy contract with a generator. To me, maybe today it doesn’t make any difference to increase the hashrate in terms of costs. But that doesn’t change the fact that in general, more hashrate implies more costs.
Had one additional thought about spaceship analogies. What we are hoping to see right now is just get that first ship into orbit. If it can just work pretty good for a relatively short time then we got a space race. People will see the potential and say oh I have a slightly better design. After a while of this bam at least SOMEONE made it to the moon.
This is completely new tech and I think its maybe a bit naive to assume the first one of it will be the perfect one we use forever. SAFE has the advantage of being first, in fact the creator, of this particular market. Later “it sorta works fine for a while” won’t be good enough to break into the market. Right now there is no competitive pressure to be perfect like there will be in a more mature market for this type of tech.
Yes, fixed cost matters also, but in this case, we can assume that:
being able to gather data, keep them (stay online all the time) and serve little GETs = fixed cost
being able to serve many times more GETs = flexible cost
Can we accept that? If yes, the question of @Sotros25 remains valid - are extensive GETs (variable cost) negligible cost relative to fixed cost?
For me this point is far from clear.
That’s a good point. I think there passionate Vaults, those who work for the good of the system and love the project. Others will join the party for fun or for profit. Thinking about a big network, everyone will have it’s own threshold to leave. Difficult to predict. At the beginning the passionate Vaults will be the majority and things will be safe. After that, things starts to get out of control, for good or for bad.
Each project faces it’s own barriers and oportunities. The better your model, the better the chances to success.
In many cases this is true. But in the case of Safe Network it’s very difficult to decouple things, at least for me.
This project is very complex, a lot of details and efforts are being devoted to different issues, algorithms are revisited, meanwhile I’m getting the impression that the economic model doesn’t matter much, it will start and then we’ll see what happens. Some here are aware of the risks, others aren’t. Changing things along the way is not so simple. It depends what you are going to change and how. Without seeing proposals it is difficult to consider that everything will end well.
I assume you are aware vaults will also be given Safecoin in proportion to their adulthood, eg. their ability to stay online for extended periods of time and serve data properly, eg. being loyal. Is that not a convincing point for you to believe it will be stable in game theory to stay even in harsh times?
I agree. On top of that, if more Vaults joining the party, the rewards will be divided…so your fixed cost can have a different weigh in your balance.
I am thinking of fixed cost in that case as kind of stable in time… Imagine, once you are an adult, the network wants (incentivise / accepts) new vaults only when they are needed. There will be nothing like a sudden influx of unnecessary vaults and shortage of work for old ones… Did I understand you correctly?
Not enough. Look Cardano for example. To maintain a pool you need some technical knowledge, keep everything online and so on. But the k factor they started was based on a scenario that is far from the real number of pools they have. A lot of operators already start complaining and will realize they aren’t able to profit. So they will leave, sooner or later. Then things will stabilize again after the hype.
Does not matter… also technical knowledge is in fixed cost regardless in that sence because it does not matter how much GETs you are getting. Back to the point - we need to know if more GETs are negligible extra cost or not. Don’t you agree?
If GETs are not a bottleneck even if there is 1000x of them (1K vs 1M for example), I would probably agree with following statement.
That is the point for me. If the network needs more Vaults, it should encourage more rewards for GETs, not for PUTs.
In some scenario, maybe the network is fine but the number of PUTs are increasing a lot, making it more profitable for the Vaults when it shouldn’t, so more Vaults will join. That’s not the worse case, the worse case is when the GETs are increasing a lot, you need more Vaults but doesn’t have the right incentives for that.
Yes, my answer is that GETs can be a bottleneck. No one have infinity GETs to offer. I am not seeing this facility to deliver more resources as easily if the network requires it.
We need to test IF the GETs are bottleneck and how big is the problem… IF yes, THAN we SHOULD continue this discussion
if GETs can bottleneck the the network is very vulnerable to being attacked by someone just requesting lots of GETs they really arn’t using. So I think there either has to be some cost to requesting data, or really GETs just don’t cost anything noticeable to produce more.
Again this is the first experiment! If it blows up no big worry. We will just restart/fork the network and try new parameters.
There’s been a lot of serious thoughts by many community members. Some have even made simulations. Here’s an example by @oetyng: Exploration of a live network economy - #56 by oetyng. There’s a lot more like it if you search.
I work for a local ISP in Czech republic so I can give you some real life numbers how people use internet. For exapmle one site we have has over 300 homes mostly with 25/5 Mbit plan. In the highest traffic hour the real traffic of this site peaks at 350/50 Mbit, that is only 1,15/0,15 Mbit per one home in average.
This means even with this not ideal scenario, if anyone of them was running vault, he could (statistically) serve 4 people with zero addional cost (not paying higher plan, difference in pc power consumption would be hardly noticeable). With 100Mbit upload I could serve 80+ homes at the same time from my single home PC.
Movies in 4k nad similar things can take a lot of resources, but dont forget most of the PCs and network connections are idle 99% of the time. (Me writing this post is idle from resources point of view.)
Even in some kind succesful of GET attack on the network, it wont damage stored data it will be DDoS like others that happen in todays interenet. Such attack costs money and what do you get in return?