At least with Google’s current business model, they make money the more people use the internet. So if the internet were to continue like it is today, then yes I think they’ll prevent data caps. If the SafeNetwork takes over, I guess that may be a different story. But I don’t expect Google will just roll over and die. They are in the business of boosting people’s internet usage, and if they can jump on the SafeNetwork early enough they’d not have incentive to implement data caps.
Good reasoning - for the most part I agree, but with SafeNetwork - at least the way it’s going to work in v1.0 - you don’t get to show ads (unless you insert them as JavaScript from the Web, which is very easily blockable), and you don’t get to log any user data (unless you also interact with the user via the Web, which is very easily blockable), so it’s going to be fun to see how exactly are they going to benefit from all that traffic.
Do you consider that for any particular ISP that SAFE traffic will be significant compared with say Netflix and the other streaming (youtube etc) services. And if streaming services move onto SAFE (would they though?) will the other SAFE traffic be enough for ISPs to change.
In Australia Netflix just started up and already they are dominating the bandwidth (20-40%). Some ISPs here installed Netflix servers at their core switching centres. Netflix also installed servers at the core exchange centres here too.
Also could we find that with caching and the randomness of location we will see that a significant (say 10-20%) amount of SAFE traffic will be within a local ISP?
Basically I am asking that in the next few years before SAFE takes over the net, will there be enough traffic to force a change in ISP’s thinking? There is a lot of traffic that will not move onto safe, or be resistance to the idea. For example major streaming, bittorrent, usenet which account for the majority of traffic now.
1 Like
In a move that is sure to spark a debate about “net neutrality” in Australia, Optus is soon likely to ask internet video streaming services such as Netflix to pay a fee to ensure that quality video streams are delivered to internet subscribers.
(Optus wants Netflix to pay up to ensure quality video streaming).
I think only a foolish management wouldn’t ask Netflix to pay up.
In case of Netflix, there’s Netflix Inc. who can give them a cut.
Who’s going to pay telcos a cut from MaidSafe traffic?
Also, people mistakenly assume that telcos simply must continue providing existing services under existing T&C’s. They don’t.
The question is do the telco’s and ISP’s charge what they charge because it reflects their actual costs? Or do they charge what they charge because they are a government instituted monopoly that can charge nearly whatever they choose to charge…
A good successful meshnet may prove them all to be irrelevant… Time will tell.
Generally I don’t trust that Government Sanctioned Oligopolies deserve much ear when they appeal on “free market” principles. “No, it isn’t a free market”
With that said there is a cost and it ought to be paid by the user. Probably through micro micropayments in a cryptocurrency…
We don’t have net neutrality in AU, although most ISPs basically have keep it neutral because it works best for them and when they are found out shaping protocols, they are hit hard by their customers. Optus slowed usenet at one time and the customers switched ports and slammed the network until optus relented.
Optus typical chest beating to divert attention from their network problems even before netflix started 2 months ago. This was a statement made before netflix had really started. They did a deal with netflix to provide quota free access to netflix, so in fact while they cry about performance they welcomed netflix with open arms. Also Optus is providing 6 months free access to netflix for their (new?) customers so I think it was more to do with recovering those costs with the threat to netflix that otherwise netflix would be paying them always.
Confusion, yes its the typical way to prevent the real facts coming out. But knowing the industry in AU this is simply oiling the request for compensation for providing 6 months free access to netflix. Also a bit rich since netflix traffic is nearly all local traffic since netflix installed servers in Optus switching centres
The (print) Media in AU is also controlled by a media company that is supplying a streaming product in direct competition to netflix and is spreading fake rumours to discredit netflix’s entry to the market in AU
1 Like
So if someone codes a script, that reboots the vault until it “finds” relatively popular data then stops and farms that, would that be feasible? Or can someone run a few vaults discover which chunks are popular fire up a fresh vault and copy those selected chunks over and farm that 
These are all the type of things people will try, I suspect whilst all the spamming is happening they could have earned the safecoin they are chasing so hard 
I am 100% sure folks will do this kind of thing though, it’s just human to try and game systems, the maximum out for minimum in approach (all the game theory stuff). I think and hope this network will let them spam away happily and reward the normal users higher 
4 Likes
Remember that files are broken into small chunks and spread very broadly, randomly across the network. There really is no way to “find” popular chunks. Your Vault holds them securely till they’re called for and then gives them over. If the chunk is VERY popular, opportunistic caching will mean that fewer and fewer requests make it back to the Vault holding the data because the requests get filled out of caching along the route. (See “opportunistic caching” definition I compiled here. (The Glossary is a work in progress, but all are welcome to consult it and make suggestions if you want.)
I don’t think there’s any gaming the system here. Turning your Vault off just means that you have to take time to accumulate chunks again, never knowing which ones will be called for. Everyone profits most by just letting the Vault do its job as consistently as possible.
2 Likes
Exactly, it’s like the ant asking while foraging, “if I find the food can I keep it for myself and not give any back the the other workers!” It’s amazing we have lasted to long as a species when you see the lengths we go to for personal greed (I am not saying @Ricmaric is in this category in any way). It’s a thing that amazes me actually, means we have to design much more into the system which I suppose is good, but we could be so much better as a race working together. I wonder one day if education will get us to the level of an ant or gnat?
Then again ants and gnats do not have money!
4 Likes
A human gnat would be a vampire.
Just sayin’ 
4 Likes
They just don’t know what they’re missing or perhaps they would.
1 Like
In addition to what @dirvine and @fergish stated… they also need to consider the Pay-Per-GET, which is based on the Sigmoid Curve.
Finding a popular chunk is not enough if you’re well below the Network Average. Your chunk may be popular but the Pay-Per-GET will be very low.
To put it in number terms, it looks like this.
Vault 1 (Normal)
Has 100GB stored with 2GB (popular chunks)
Vault 2 (Scripted)
Has 10GB stored with 2GB (popular chunks)
If the Network Average is 50GB, Vault 1 will earn a lot more Safecoin per GET compared to Vault 2, simply because it has more chunks stored. Even chunks that are not popular add to the vaults earning potential.
4 Likes
5 Likes
Well, ants dont eat the leaves but the fungus that grows inside the anthill; the leaves they collect are brought to the anthill to feed the fungus, so they don’t have the luxury of thinking selfishly because they can’t eat what they individually collect.
So they have no choice actually but to work collectively to survive: their selfish interest of survival leads them to the only feeding mechanism that benefits the community as a whole.
It is like having one single kitchen, to eat myself I would end up cooking for everybody else as well.
It is either that or eating raw food, which we can’t digest.
I think this functional duality of everybody safe or nobody safe -by design- is clearly reflected in MaidSafe.
2 Likes
@dirvine what do you think about this,the user could choose:
-
Spend a lot of bandwidth. That is bad for the user but safer for the network because network are not pendant of that node anymore
-
Spend a few safecoins and avoid that:
If i want to reboot my computer I send a notification to the network, the network send me a new ID based on a checksum of my current chunks and stops the transfers between me and the network. After that i wil have X minutes to re login myself and that could be possible ONLY if my checksum is the same and then validate my new private keys (proof of burn?). If the checksum don’t works so he must re-fill their vault. If the user does not reloggin properly in time he are going to get bad reputation because he promises something false to the network.
2 Likes
Sounds Interesting, I though need several days/weeks to consider these things, so fall back on the community to debate and then hopefully when an RFC is created, if approved, we can see it in code and then create simulations and look for edge effects.
Sounds like something interesting though for sure, first get the logic in place then add human greed and mental vandalism attacks though. Takes a lot, but the more in depth debate the better for sure.
Sorry I cannot spend a lot of time on all the ideas recently (been an awful lot ) Great though and hopefully with the bounties being paid this will add even more in depth ideas. XOR is very very hard though, understanding attacks is even harder and takes ages.
PS: This could also link to archive nodes that will require to hold data consistently over longer periods (perhaps)
2 Likes
Wondering why SAFEcoins would have to be spent. Why not just allow vaults to “ask” the network and the process as you outlined occurs. The whole idea is to farm and be rewarded SAFEcoins, but if its going to cost a significant portion of my rewards to save the refilling of the vault, then is it economically viable? Especially when there may not be a network/security benefit to spending the coins?
Sorry for post twice. i cant delete my reply to irvine.
Wondering why SAFEcoins would have to be spent.
Of course should be not very expensive.
The idea of paying safecoins to avoid refill vaults (and wasting bandwidth) have two intentions:
-
Discourage desertion because refillings takes too much, in a responsible way (all my gamers friends are potential farmers for example, gamers needs reboot)
-
Make not economically an attack where someone lies to the network about a temporary restart of X vaults simultaneously and then not turning on them in order to delay the reconstruction of redundancy.
Anyways, the network waste resources and time checking if you are a good boy or a bad one so makes sense pays for that “work”.
1 Like
Still not sure its needed to spend any coins. It seems that a restart is a reasonable request of any vault and the nodes can allow the X seconds for the restart and likewise X seconds before allowing another from that vault. If you are going to have one time period then 2 is reasonable.
I do like your idea, just wondering if the spend is needed. So agree with this point, good idea BTW
I would expect that when a vault goes offline, even for restart, would have the network starting to “reconstruct redundancy” This is because
- restarting does not always go to plan. Sometimes things need fixing
- people (vaults) lied
- so the network does as it always does when a vault goes offline, it starts making new copy of each chunk. Remeber that if X seconds is 120 or 180 or say 600, then this is a long time when you are trying to have a redundant file system.
- If your vault comes back online then having a fifth copy of some chunks is not going to make much difference. The longer you take to come back online means there will more chunks with a fifth copy and thus successful farming will be a little more difficult
- This then is the penalty for needed a restart and I would argue for network data security it would have to happen anyhow
The extra work is in giving you a crypto signature before restart, the checking afterwards is simply that, very quick and either the network will continue your vault as it was or it will refill it with new stuff. Both would be normal operations.
Actually this can happen anyhow, a city like Sydney (~8 million people) suffers severe storms and power is lost, imagine the 1000’s of vaults offline at once. Guess I am saying the system handles this situation standing on its head, Doubt someone lying about restarts affecting a number of vaults is going to be noticeable, they lose rewards, bandwidth of filling their vaults, etc if they do it often. My guess is they get sick of it, especially when the network is designed to handle large number of vaults going offline at once. The only effort is the crypto signatures, but when you consider that amount of work the system does anyhow this is really a bump compared to that.
The crypto signature is not going to take as much work as you might think
- the vault does the all the hard work
- then asks for a signature based on that
- If the vault lied about what was in its vault then normal system processing will downgrade it or invalidate it later on when chunks are retrieved and they are not correct. The system only knows a vault is working correctly when the chunk is retrieved.
- the system cannot create any “checksum” of a vault because the system does not know the chunks and has no way apart from asking the vault. The proof of a vault working properly is that it gives a requested chunk that matches the signature used to retrieve it. (I hope this last part is completely correct, but even not the principle of getting the vault to create a checksum is valid)
1 Like