Feature - 2 Factor Authentication

jpr said:
I think this is going to be built into darkcoin protocol. No centralization here :)
Click to expand...

I got triggered by the last line of Evans proposal, it read to me like some external service will be used:

More research must be done to find a compatible 2FA API. There are many services to choose from and we'll evaluate each to find the best match.

What is meant here? Fork an existing solution into Darkcoin?
 
As Darkchild mentioned a bit ago, FIDO seems to hold promise. I have really appreciated my Yubikey, and the freedom it gives me to walk into an Internet cafe and not have to worry about my passwords being sniffed. On the other hand, it is something of a concern that their servers must be operational to authenticate my device, even though they cannot see my interactions. It looks like FIDO avoids that problem, though I will need to understand it better to be convinced.
I suspect that the master node network would be capable of serving the same service using FIDO and this device: https://www.yubico.com/products/yubikey-hardware/fido-u2f-security-key/

As FIDO is an open standard, I suspect its implementation would be almost trivial for someone who knew what they were doing, and that any number of similar devices will be available.
 
Just saw this. I'm a little amused that this was posted 3 days after I lost everything in my own Darkcoin wallet - I wonder if I was any part of that xD

Fantastic work Evan. If you pull this off... I can't imagine how game-changing that would be.
 
MaxFangX said:
Just saw this. I'm a little amused that this was posted 3 days after I lost everything in my own Darkcoin wallet - I wonder if I was any part of that xD

Fantastic work Evan. If you pull this off... I can't imagine how game-changing that would be.
Click to expand...

''would''…. When… :)

Sorry to hear about your DRK, if you don't mind the question did your password get compromised or did you not have backups and your computer data got corrupted? I hope you managed to recuperate your coin.
 
It's the most simple and obvious ideas that always seem the smartest. You'd think that would make it easy to come up with them!
If I understand it right, this would/could act as a sort of Darkcoin savings account, or maybe more of a vault (Dark Vault? Sounds like a SyFy Original), and I'm really all for that idea, even if the idea is only in my head due to a misconception.
 
I do not think that this method (fido) is good. We need something like a second key with whom we sign transaction.
Let me explain how it should look like (IMHO):
- in wallet I make a transaction
- on the screen I can see 2D code for this transaction
- I scan this with my phone camera with speciall app
- app sign transactions with second private key
- on the phone screen I can see 2D code
- with my notebook camera and wallet app I scan this
- all is propagated to DRK network
No central servers or other stuff.
 
Thanks for explaining and sorry to hear that! You definitely need a password and then you have to watch out for software that copies your password so you should maintain a good security level overall… I also had a hack attempt many months back.. I hope you got yourself some nice new fresh DRK and that you are ready to make your money back many times fold :)
 
strix said:
As Darkchild mentioned a bit ago, FIDO seems to hold promise. I have really appreciated my Yubikey, and the freedom it gives me to walk into an Internet cafe and not have to worry about my passwords being sniffed. On the other hand, it is something of a concern that their servers must be operational to authenticate my device, even though they cannot see my interactions. It looks like FIDO avoids that problem, though I will need to understand it better to be convinced.
I suspect that the master node network would be capable of serving the same service using FIDO and this device: https://www.yubico.com/products/yubikey-hardware/fido-u2f-security-key/

As FIDO is an open standard, I suspect its implementation would be almost trivial for someone who knew what they were doing, and that any number of similar devices will be available.
Click to expand...

Actually, you can configure a yubikey to generate a one time password (OTP) in a similar process to google authenticator. I'm currently playing with the yubikey, and while I won't say it is the perfect solution, I think the idea of having the option of adding 2FA should be based on something like the yubikey or Authenticator.

Simply logging in to your wallet and having that login be verified by a third party server creates a time signature that could then be matched to any transactions you make. although Darksend mitigates this to an extent, if I know you logged on to your wallet at 0711 UTC and then I see a bunch of darksend transactions for the next 4 minutes, I can make an assumption that one of those is yours. Instead of trying to sort out darksend transaction, I can then look at transactions to known entities, like exchanges, and see if any of those match.
This is all highly theoretical, but why create a potential vulnerability?

IMHO the 2FA should be something you ACTUALLY have, not something a third party has. And as always, I'm a huge advocate of flexibility in enabling users to manage their own anonymity as much as possible.
 
HammerHedd said:
Actually, you can configure a yubikey to generate a one time password (OTP) in a similar process to google authenticator. I'm currently playing with the yubikey, and while I won't say it is the perfect solution, I think the idea of having the option of adding 2FA should be based on something like the yubikey or Authenticator.

Simply logging in to your wallet and having that login be verified by a third party server creates a time signature that could then be matched to any transactions you make. although Darksend mitigates this to an extent, if I know you logged on to your wallet at 0711 UTC and then I see a bunch of darksend transactions for the next 4 minutes, I can make an assumption that one of those is yours. Instead of trying to sort out darksend transaction, I can then look at transactions to known entities, like exchanges, and see if any of those match.
This is all highly theoretical, but why create a potential vulnerability?

IMHO the 2FA should be something you ACTUALLY have, not something a third party has. And as always, I'm a huge advocate of flexibility in enabling users to manage their own anonymity as much as possible.
Click to expand...
I whole-heartedly agree, HammerHedd! The "something you know-something you have" model of security is the way to go (IMHO).

I too appreciate the OTP approach that yubikey has implemented, and like you, have some doubts about its direct implementation into DRK. In addition to the potential of timing as an attack vector, I am also concerned that Yubico's OTP implementation seems to require the registration of a particular device in generating the OTP. While I don't understand it fully, I suspect that this would allow the linking of a users accounts, even though third party access to those accounts would be impossible. In other words, If I use my yubikey to access my Gmail acct, and the same device with another key for DRK, the two accts could be identified as having the same owner, even though the transactions themselves would remain secure.

My reading of the FIDO standard makes me think this would NOT be the case with it, but I would want that confirmed by others more knowledgeable than I. I would like to think that an implementation of FIDO in which the wallet requires the password to be opened, followed by entry of an OTP confirmed by the MasterNode network prior to broadcasting a transaction, would be both secure and feasible. I would also hope that the FIDO standard would allow the printing of OTPs for emergency backup (as implemented in the yubikey) would also be possible.

While droning on I would also add that I can envision a system in which miners must register with the MasterNode network in a similar manner prior to block acceptance, and that this mechanism could be leveraged to provide protection from the 51% pool dominance that so many of us are concerned about.
 
I just read the abstract, and this does look good. However, I hope that any implementation we make of 2FA for DRK will not be dependent on using a cell phone, as I suspect this would clearly identify the cell owner as a DRK user. At this point, in most jurisdictions, this is not a problem, but I have not yet read next month's newspaper.

The risk would be mitigated once the MN's identities/IPs were obscured; provided they sent the 2FA code in a secure manner. Whatever method is ultimately chosen, I hope it will use FIDO in order to be able to use a Yubikey-like fob.
 
Not to throw cold water on the 2Fa discussion.. but shouldn't we implement instantX first?

I thought since Evan had a POC already working for instantX that implementation on testnet was imminent.
 
jimbit said:
Not to throw cold water on the 2Fa discussion.. but shouldn't we implement instantX first?

I thought since Evan had a POC already working for instantX that implementation on testnet was imminent.
Click to expand...
I think InstantX is probably going to be next for testing. This announcement about 2Fa was like weeks ago, the news just now caught up with it.
 
2fa is all good, but do keep in mind those who do not use google or apple and may need other options. So many sites assume authy, etc. as standard 2fa.
 
You must log in or register to reply here.
Back
Top