Seems like such a major and obvious breach in protocol, that I wouldn’t have even thought this was possible to do and have the network pay for it. Would love to hear from the team if this is indeed possible, and not being worked on to be fixed before testcoins come around.
1 Like
It is how all chunks are retrieved - by address. The datamap is a list of chunk addresses that are retrieved from the network to read a file. Its just how the network works. Even have a URL format to access by chunk address. Its a core aspect of the network.
It is the client code running on your computer that does all the assembly and decryption of files.
Now in the gaming of farming there is no need to decrypt the chunk, just request it.
1 Like
I just meant that I didn’t think a specifc vault could request a chunk from a specified vault… and especially constantly…
Yes correct there are eight chunks but you could modify your node’s code too.
Now since you are closest to your node then yours to be the preferred vault has a head start over the others. And if you request 100 times there is a good chance your vault will be the one rewarded in a good percentage
You make it sound a bit too easily gameable, and I don’t know all the technicalities, so I would like to see someone like @dirvine 's opinion on gaming the farming to that degree…
But at any rate we are off topic in this thread, the question being should PtP be implemented?
I say Yes.
Should it be implemented in the core of the network?
I say definitely not, because it is far more easily gameable than farming seems to be.
And really I’d like to see some good in depth explaination as to why it NEEDS to be in the core of the network as opposed to a seperated second layer type thing, or an app (possibly installed and enabled by default that people can turn off), or preferably even a seperate/secondary coin.
8 Likes
I have not read the whole thread, so don’t know 
The farm algorithm we will start with is very naive and that is on purpose, for many reasons. As we gather data and try ourselves as a community to “game” it then we will see if there are adjustments to the algorithm and vault rules that need to be made. May seem weird, but we have Engineers working in al areas of this network right now and I want that all to come to a head with a minimum feature set for a running network. then I would like to see these conversations polluted with maidsafe Engineers as well, but for now we don’t have the time allocated for that. We will soon though, all going well.
14 Likes
Yes, a possibility right now. Just imagine how far along we’d be if this thread had 622 posts of brainstorming about threat mitigations.I propose we focus the thought experiments there.
Agreed. Your posts raise the pertinent issues. This is an opportunity for some brainstorming here, which costs the developer team nothing and can possible lead to some good ideas on how to protect farmers and Dev rewards too.
3 Likes
I would suggest there to be a reservation of a percentage of the network income and there be a secondary system where people get to “tip” someone and each month or each day the reserved percentage of income is distributed to all the people that have been tipped!
so some kind of tax for every operation that costs safecoin that is being given to wallets that have been tipped by other users.
so e.g.:
- the network get paid 1000safecoin for people to upload stuff,
- then the 200safecoin would be reserved.
- then people would tip some one or some content lets say in safetipcoin. so one content uploader would get tipped by 3 people at 10 safetipcoins and another uploader would get tipped by 4 people at 6 safetipcoins then the first content uploader would get 125 safecoins from the 200 reserved → (sum of safetipcoin 16, first one is (10/16)*200 = 125)
and second uploader would get 75 safecoins → (6/16)*200 = 75
how about that?
Not weird at all imo - I find it very reassuring! I can’t wait to see the birth of a working network! There is plenty of time for it to mature later.
6 Likes
Why wouldn’t a content uploader simply tip themselves with some other accounts to garner some of that supply? That actually seems easier to game.
2 Likes
everyone will have only some “voting” “tipping” points
edit: ow now I see your point, well why wouldnt someone make many vaults to make money? its the same thing
maybe to combat this, tipping points are earned by vaults or bought
All the loot should go to computer programmers because they are like the most high…all others should eat cake.
1 Like
Has a design been found for PtP that isn’t susceptible to an exploit where I upload some data then set up a botnet to request it in a tight loop and siphon tokens?
3 Likes
Not really @anon57419684 there is no final design as such only an outline, however, to be based on use, so those attacks are valid to prevent. If we only use get then it can happen, but if payments are based on mutations it’s better (where payments are less than the cost of put/mutate). It’s an intricate issue though.
5 Likes
My fear is that it will encourage DDoS that wouldn’t otherwise be desirable or affordable. Spamming the network with requests from many different places may get a hit on the vault and avoid caching. However, this is the exact opposite from normal user behaviour, where they should get cache hits.
3 Likes
Cost benefit. Look at the cost of doing botnets. There is a reason DDOS attacks only last hours to a day or two. PtP from my understanding is not a huge money maker in short periods. Caching messes that up.
The problem as I see it is that setting up such a BotNet would be extremely difficult when you know where the correct nodes are to hit to avoid caching,
But with Safe that is impossible because of 2 major reasons
- the route the data takes is unknown to the attacker.
- the attacker does not know which node each bot will connect to before it connects.
To overcome these issues becomes a high cost for the Botnet and reduces its effectiveness and time it can actually achieve anything before it dies off.
People seem to think that PtP will pay out big time. If it did then its unsustainable longer term as more and more works are added to the network.
Caching will mean that the rewards will be paid out on a more organic basis with millions of people around the world accessing good content rather than 10K or 100K or 1 million bot operating for few hours up to 2 days getting defeated by caching.
TL;DR it needs to be tested with caching turned on.
Yes there will be some income made from it. Question is if it will be greater than the cost of time, effort and dollars to set up the botnet to do it.
Every system will be gamed to some extent. But is that a reason not to do it? Testing will answer this for PtP.
For instance farming will be gamed. People will spam malware out to 100’s of millions of email addresses to try and install a node on the poor person’s machine. Honestly this would have the potential to make a lot more money than any botnet to try and get pennies from PtP
EDIT: The huge benefit to the network with PtP is the attraction of people to the network and to upload their favourite tutorials, screenshots, videos (eg youtube replacement) and so on. Also it will encourage artists to upload the best quality to the network and so on.
This will be a valuable aid, in my opinion, to acceptance of the network by artists and people with a lot of content people want to view. More people viewing means greater network effect and so on.
So even with people maybe gaming 10% (won’t be possible in my opinion) of all PtP rewards the results from having it could mean the network accelerates in acceptance and adoption. Also helps against the economic attacks Dimitar talks of because its one thing an attacking player would include in their network.
4 Likes
I assume from above that if I upload an amazingly popular video, the system logs that I was the uploader and I get paid. I assume then that if someone downloads ‘my’ video, and adds 1 second, and then uploads this new version to the network, then the network will think it’s a new file and theoretically speaking the revenue may be split between my original and the altered copy - assuming the new version gets equal viability? This is probably going off-topic but interesting to know.
1 Like
Like PtD, PtP would require a coin balance ID to pay the rewards into.
Yes
That is why artists would want to have brand acceptance. So if an artist uploads the best quality of their works (and other qualities perhaps) then people will naturally go to the actual artist rather than some knock off that must have changed something. I am sure people would find ways to build up their “brands”. One other example would be the safetube where people go to a person’s channel because they enjoy it. So copies might get some views but I’d expect nothing like the original channel.
2 Likes
It the system is chunk based for public immutable data they would need to make a lot more changes than that. It would require manipulation or alternate encryption of every chunk that makes up your work. Original artists could embed an id to show which works are authentic.
3 Likes