Once deployed, how flexible is the SAFE network?


One question keeps popping up in my mind (and it’s 3:30 am). Better get this off my chest… :wink:

Once the full network is deployed and millions (hopefully billions ;)) of users are using it…
how difficult will it be to make changes to the network?
I mean fundamental changes like switching to another encryption mechanism or
incorporating some new kind of service/technology that comes along…

And who would be making the decisions of what and how to change it? would full consensus from
the whole network be required? would there be some kind of vote? could that lead
to forking of the network?

I think it’s important that the network is not static but able to keep evolving. But how exactly
would that work?



This is something we do put some effort into. The current encryption schemes used are templated in the code to make a change easier. So if we say moved from RSA signature to a lattice based system (quantum secure) then we can adjust the template. In this case each signature for an upgrade would be a double check, one lattice and onr RSA. This allows vaults and clients to live with two schemes per upgrade.

This is one of the reasons vaults will auto update as we cannot afford to let them get old. As its a decentralised autonomous system vault managers can actually test upgrades to ensure they equal or improve the current system (to prevent a hack by replacing vaults with bad nodes etc.).

For client apps then it should also be similar, but as we take care of all the encryption and signatures under the API to allow app builders to not have to be crypto aware then that will also help.

So not simple, but yes under constant review.


Is there somewhere we can read more on this and how its made secure?


Not yet, this in the code unfortunately still no papers on it though.