I would think so, not aware of the details though.
The argument is that several chicken eggs in a basket is safer than one ostrich egg in the same basket.
I hope people do split vaults as it adds to the number of good nodes on the network (safety in numbers does work in decentralised systems).
Doesn’t this assume that only the good guys will split their vaults? I believe it will be the opposite: splitting the vaults among multiple farm instances artificially increases the number of “good” nodes at the slight expense of performance (because of less efficient use of compute resources).
If I were the bad guys I’d use the same approach, but I’d take it to the extreme because I wouldn’t care about being in the farming business anyway.
If it makes sense for the good guys to run up to 4 farms per HDD, there’s no reason why the bad guys wouldn’t run 40 per HDD. At the very least they too would run 4 farms per HDD so overall effect on the network would in the end be none. (The little guy with ARM-based box who can barely run 1 would probably be hurt).
Read and write requests are not supposed to be concentrated on any individual instance and you should assume that if an instance is busy there’s a reason for that (like, it’s making SAFE coins for you). The main problem with this approach is a small performance hit and more labor-intensive maintenance.
Let’s wait and revisit this once the beta is out!
When the inventor of SAFE comes out and declares:
I hope people do split vaults as it adds to the number of good nodes on the network (safety in numbers does work in decentralized systems).
…and someone counters with
I believe it will be the opposite
I think I’ll stick with the inventors reccomendation.
2 Likes
Running on belief, eh?
The reco applies to the bad guys, too, and as such it doesn’t add anything that didn’t exist before it.
In fact it is probably going to be more appealing to the bad guys to run multiple instances because given the same resources it disproportionately increases their chances of success. At the same time the good guys have to use extra effort to manage and run multiple instances of MaidSafe for what is essentially a negative economic gain compared to running a single instance (or the economically justified minimum number of instances, if they have a lot of storage space).
But hey, whatever floats your boat…
@janitor you’ve just explained why more good nodes (good nodes spitting) benefits the network - which shows the error in your egg analogy.
In addition, I think @dirvine just said that the network incentivises good vaults splitting by penalising size, which mitigates against your new point about efficiency incentivising good vaults to not split.
BTW Your constant sarcastic or superior tone doesn’t add weight to your arguments, for me it’s the contrary.
2 Likes
All right, I’ll tone it down. I turn sarcastic when I spot the absence of arguments.
Good, at least someone recognizes that my posts contains arguments 
I meant to say while the scenario is indeed beneficial, it is not sustainable.
Before you posted this I went to “Off Topic” area to start a topic on farmer centralization. I claim that centralization won’t be economically feasible. But within that group of big pro-farmers there may be various ways to be relatively more efficient compared to one’s peers.
My critique was related only to the the idea that there is economic justification for running more vaults than the minimum number justified by each setup (say, if I have 48 HDDs in the box and 1 CPU with 8 cores, I may run 8 MaidSafe instances; I would not want to run 48 MaidSafe instances just because it “helps the network” - that’s the expectation that I said is unrealistic and can be replicated by the bad guys too to level the playing field to the original state).
1 Like
I think most people on this forum are not about arguing who’s right or wrong. They just wanna learn about this MAIDsafe concept. It’s quite new, very hard to understand and just as hard to explain for people like David who are very deep into it. A number of times you referred to older posts where “things like that” were already discussed. Well, if so, provide a link to help people out. I’m quite an active reader on this forum and even search doesn’t provide all the answers cause people use different names for the same thing. One could ask about a node, while the other already gave an answer two months ago referring to a vault. That has nothing to do with people not using search.
2 Likes
I sometimes browse this forum on iPad and for those who haven’t tried, it’s a nightmare to open another tab, search for stuff and paste the link over. If you try you’ll see how replying to a post with 2-3 links can easily take 15 minutes.
Well, think about it this way: you could read the entire forum and you’d find the post. Of course that’s a bit crazy and no one wants to spend that time, so some search 5 minutes, others 30, and yet others (or these same guys, if they don’t find answers) just create a new post.
No one has to reply to any post here, but in absence of other info the value of knowing that something was discussed is that if you spent 5 minutes searching now you can - if you don’t want to wait for more useful replies - give it another try and find that content.
Here’s two topics related to recognizing (existing) chunks on the network:
The first is https://forum.autonomi.community/t/recognising-file-copies/156/8.
The second is below and it was created by you. You asked almost the same question yesterday and before (and that time @dirvine answered it).
- Yesterday (scroll above or Grouping vaults by owner, to strengthen redundancy? - #6 by anon40790172)
- Back in August: What about nodes sniffing data? - #3 by dirvine
Thnx for the links. But these are 2 different questions. The question about “nodes sniffing data” is about a user (node) finding out which chunks are requested by other users (nodes). The question yesterday was about me (on my own node) finding out which chunks are on my computer in my own vault. That’s a very big difference and a completely different question. So your argument doesn’t hold.
There’s an argument to be made that helping the network helps you, because you profit from the network staying in good health (and thus attractive to users).
Depending on how your instances are set up, you might be able to marginally increase security by splitting them as well; an attacker would have to break into them one by one to compromise all your vaults, giving you some extra time to cut them off.
Another benefit for the farmer would be the ability to add more hardware and get a head start on rating with it by moving one or more of your existing vaults over and increasing their size, rather than creating a new one.
I sometimes browse this forum on iPad and for those who haven’t tried, it’s a nightmare to open another tab, search for stuff and paste the link over. If you try you’ll see how replying to a post with 2-3 links can easily take 15 minutes.
It is a pain, i agree, and it is also hard to track down posts using search, even on a desktop screen! I’m almost always using a tiny smartphone (Samsung mini) but still try to include links and quotes because it makes my posts useful, and that is the point of posting for me.
A lot of posts, including some of mine, suggest people search, but they are one tenth as helpful as a link, so all of us should bear that in mind.
And thanks for taking my comment on style so well. I’m prone to similar foibles so feel free to point it out - “if you can spot it, you got it” applies :-). You have a keen mind and I’ve learned much from you. We all have our styles, so I try to read past style, but I often find it hard to read past yours and I’m sure I miss some value.
I haven’t read the other post you mention, but it doesn’t sound off-topic. It sounds like a Network issue, so maybe re categorise it?
1 Like