I think the Flemming features changed on safenetwork.tech 
https://safenetwork.tech/roadmap/
10 Likes
And all marked as implementation phase 
8 Likes
So I would speculate on mid to end november release… 
9 Likes
That looks like a full network.
9 Likes
Team is going ahead as fast as never before. Wow that’s amazing 
12 Likes
6 Likes
First version of multiple local vaults (mock routing) is done!
https://github.com/maidsafe/safe_vault/issues/853#event-2720346604
13 Likes
It’s going very fast now, I am very optimist again
3 Likes
There is one point on the roadmap which I don’t fully understand:
Obfuscating stored data
Isn’t this already done with chunks/self encryption?
1 Like
For files yes. But for AD objects it may or may not be encrypted by the user.
Also if you self encrypt a file then you can identify chunks in your vault if any of that file is stored in your vault.
This point will make the data obfuscated for the vault operator.
6 Likes
I thought that on the chunk level the data is always encrypted? Vaults are currently storing plain text chunks in case of AD?
The obfuscation part makes somehow sense to me now.
That is absolutely mind-blowing, if it is working as planned
1 Like
Chunks are for immutable data and is encrypted via self-encryption. Remember the user is doing the encryption and why its called self-encryption.
Mutable data (ADs etc) are not enforced encrypted to allow public data to exist since there is no self-encryption for that data. Some mutable data though is encrypted by the user.
Public mutable data is store unencrypted because there is no standard (self) encryption that the public can use to see that public data in the mutable objects. It is possible that the keys could be stored with the data, but if its public then thats wasted work for the most part.
Private mutable data should be encrypted by the user before storing otherwise its public data when found.
Now back to immutable data. Lets say I want to find if a file is stored in my vault. So I self-encrypt the file and take note of the addresses generated. I then discard the chunks generated since I don’t actually want to store the file
Then I check the chunks stored in my vault and look for a match of the hash. I cannot decrypt the chunk but I know the hash and so I can compare the hashes in my vault with the hashes I noted previously.
Thus I can see if any of that file is being stored in my vault.
Imagine having a database of chunk hashes from 1000’s of big files (maybe a group effort to build) then I can know if a vault is storing any of those files. Imagine that group being the NSA, or SONY, they could then run 10,000s of vaults and see if any of those chunks appear in any of their vaults. They might not be able to do much or they could attempt further cracks of the system
Obscuring that data then prevents me from doing that trick looking for chunks and all mutable data is hidden from the vault operators.
I hope this helps you to understand that using obscuring then the data is safe from prying eyes and all part of the plan to achieve this.
5 Likes
Thanks @neo it all makes sense to me. I just thought that all data is encrypted, even public data but that this is publicly readable by the network.
Isn’t AD data also somehow stored in vault “chunks”? If not how is it stored? If I look into my vault dir there are only “chunks” that seem encrypted.
Public files are still self encrypted but since the data map is provided (the keys) then anyone can read it.
Its stored as a mutable object. As far as sending it around the network and storing it, it looks and acts like a chunk. Has the same maximum size as a chunk. But the section storing the mutable data does a lot more than just storing it. It appends data to the AD for instance. Checks ownership and permission for another.
2 Likes
So if I understand correctly everything is self encrypted but some data might be identifyable by the vault/data owner which obfuscation should solve?
1 Like
No mutable data is not self-encrypted. There is no method to do that, it not logically possible.
any encryption for mutable data is done the normal way to encrypt data. Thus it is up to the user and/or application if the data stored in mutable objects is encrypted with a key.
Obfuscation solves BOTH the out of band identifying of chunks and unencrypted mutable data that is stored.
6 Likes
If maidsafe is aiming for release asap. does this mean that the restarts and upgrades will be introduced later, after the launch?
1 Like
Why would they since they said they will implement them.
Remember beta is not launch and we have a number of versions before launch
1 Like