Nice article @nicklambert . It doesn’t address however the issue of a Sybil attack, I’m sure that anyone unfamiliar with MaidSafe but understanding of why the blockchain works will immediately think of that.
It’s impossible to thoroughly explain all of MaidSafe in one blog though, so maybe it would be a good follow-up?
Yes it’s difficult to compartmentalise aspects of the system and the reader will always want to know more. Disappearing down rabbit holes is a common problem with the SAFE Network as you know. The defence against attacks may well be a good follow up, I’ll give it some thought. Cheers for the feedback.
Has there been a decision to move from close groups of 4 to close groups of 32? It seems much more robust now.
Yes considerable testing shows this is ultra conservative and provides pretty extreme levels of security. It also allows us to get around many more NAT devices that try and prevent p2p traffic.
I think it is good for decentralized applications to be conservative about security, as in the end you cannot trust any single user to do what they are expected to do, only the whole network if it is well designed. Being conservative will help with user adoption and trust.
It is one thing to guarantee most of the data is safe under realistic assumptions, it is another to guarantee that every single chunk of it will be safe, even under extreme circumstances.
I doubt anybody would use a potentially unreliable backup system, so I would say the resource trade-off is worth it.