Trustless or bust: future of canaries

[ioptio]

Trust-less or bust: the future of canaries

Continuing the discussion from Dev Update week beginning 24 november 2014:

@ned14

The EU wide law is the Data Retention Directive. That is implemented in custom variants by each country, so it isn’t actually a law itself, rather guidelines for each country who may diverge substantially from the guidelines. The UK’s implementation is the Regulation of Investigatory Powers Act 2000 which provides very wide warrantless data retrieval powers and gagging being able to publicise the event. These powers were expanded by the Data Retention and Investigatory Powers Act 2014 which if I read it correctly, could allow the government to seize TOR nodes and operate them on your behalf for sustained periods of time. This power isn’t unusual, other European governments have already made use of it by seizing TOR nodes.

@ioptio

I’m pulling the canary because of this but there is an accompanied blog post which I still want to publish - it just needs some updating to reflect this

The canary only works because the US is very unusual in that you cannot be compelled to lie under oath. You can be compelled to be silent, but not to lie. I think that’s simply a happy accident by the founding fathers. It is definitely unusual.

However to be honest Paige a canary is a gimmick anyway. Much more valuable is making public a threat matrix which shows in a single place all the threats which can be exerted on the SAFE network by adversaries. You then aim to plug those threats.

For example, in the UK it is illegal not to supply your encryption keys to the government. If they got the Maidsafe github keys and forcibly took over our servers and gagged us from telling anyone, they could, pretending to be us, modify the code base to subvert the network and push it out as authentic updates. That is but one threat entry in the matrix. Another might be that we can’t afford our own data center so we rent them, well that leaves us wide open to governments fiddling with our servers and we’d have no idea. Or they could trap all packets entering and exiting the main internet junctions and build a pattern - this is how they subvert TOR et all.

[kirkion]

@ioptio

For example, in the UK it is illegal not to supply your encryption keys to the government. If they got the Maidsafe github keys and forcibly took over our servers and gagged us from telling anyone, they could, pretending to be us, modify the code base to subvert the network and push it out as authentic updates.

Here is an interesting question. As individual nodes, users don’t have to accept updates from Maidsafe do we? Once there is a stable implementation of Maidsafe we don’t have to update unless there is a hard fork. Right?

[dirvine]

Soon after launch this will be a non issue. I have a mechanism in the background that will work like this (well close to this)

Vault_manager gets downloads from SAFE (so crypto signed, etc.)
Manager starts a sacrificial extra vault
sacrificial vault is monitored over a randomised period
if the sacrificial vaults rank is at least equal to changes in line with existing vaults then the update is considered good

This is the basic approach, so the network will not only protect against us going rogue or bad, but also us providing code that reduces effectiveness in any way (the ultimate Qa if you like).

I am not going deep into detail here, but you can see when you consider this, of course the network is autonomous, it can be programmed to live (in fact our fallback udp/tcp port is 5483 (live if you look on your phone). This is the sort of thing truly autonomous systems can actually do. So as I said previously everything the current Internet can do is fine, but what we can now consider beyond that is another level of innovation I cannot begin to imagine, Things like updates though become very much more interesting indeed, imagine write a new node (or computation) and try and launch it, the network will tell you if its any good (its partners will de-rank bad behaviour, failures etc.)

I hope you can get a glimpse of where we can go with this

[kirkion]

@dirvine

so the network will not only protect against us going rogue or bad, but also us providing code that reduces effectiveness in any way (the ultimate Qa if you like).

@dirvine You know every time that I think I’ve heard pretty much every awesome thing about SAFE, I learn something new. This is such a supremely rational way to do updating.

Wow. Mind blown again.

[Nigel]

Oh that is brilliant! All smiles here

[ned14]

Apart from needing maybe a few more citations to support facts stated, a great article Paige.

Niall

[happybeing]

What @kirkion said.

I’ve been wondering about this for ages, and wow, just wow.

Take a bow @dirvine

[ioptio]

Thanks Niall. Re: citations you referring to the various mentioned regulations at the beginning? Anything else?