Quote from: dirvine on April 24, 2014, 11:14:54 AM

Quote from: FrictionlessCoin on April 24, 2014, 11:08:44 AM

Quote from: superresistant on April 24, 2014, 08:03:12 AM

Quote from: EvilDave on April 24, 2014, 07:51:12 AM

I have a question:

MaidSafe is, in theory, going to provide an encrypted, p2p distributed dropbox functionality, correct?

So how is MS going to provide guaranteed 24/7 access to clients data?
Farmers will provide storage space, and recieve SafeCoin as an award, but these guys are mostly not going to be IT pro's who can guarantee 99.9% uptime.
How will MS deal with a farmer that drops out, taking the stored MS data on his/her hard drive with them ?
The only option that I see to come close to guaranteeing 100% uptime of individual clients data will be to stripe/mirror that data over multiple farmers, but even that still doesn't give 100% certainty that all of the data will be online and accessible at any given moment, and will massively increase the amount of storage needed on the MS network.
Anyone have any idea how this will be implemented?

Exactly.

If you consider the fact that the competition for money will centralize all the network in few locations with the best performance and highest bandwidth (like in all crypto), you can be sure that there is a risk of losing data.

The Proof of Resource calculations appear to incentivize Quality of Service, so an unreliable node will tend to be avoided. Furthermore, there is a level of replication such that even if nodes fail, the data will always be available in other nodes. 100% guarantee however is likely not possible. I think however we are working with very low probabilities here with the data being striped and mirrored.

Excellent opinion, all of the network is a probability matrix really. I suppose all crypto is as well. The key as you rightly point out is to make the probability of loss or collision so low it is negligible over a huge timescale. It is hard to get your head around this for some people, me included at the start, as you think cpu == 100% correct math all the time. It's just not true in real life, everything is a probability and this is an excellent point, well made.

Thanks (btw address space is 2^512 so larger than the number of atoms in visible universe almost squared, not always good to use that number but a good starting point in the thought process)