Zoom Logo

Secure Data Storage - WG - Shared screen with speaker view
Dmitri Zagidulin
02:20
https://hackmd.io/fk-9A75CRkyFxodpNX3D5g
Dmitri Zagidulin
02:25
Agenda/Attendance ^
Dmitri Zagidulin
03:54
Attendance/Agenda: https://hackmd.io/fk-9A75CRkyFxodpNX3D5g
Manu Sporny
05:43
+1 thank you for showing up at 5am! :)
Dmitri Zagidulin
05:45
much appreciated, 5am!
Andreas Freund
06:06
<3
Chris Lee
06:18
haha glad to be here
Andreas Freund
06:51
3Box rocks btw!!
Orie Steele (Transmute)
12:31
@Daniel, isn’t this exactly how you wanted hubs to work?
Orie Steele (Transmute)
12:55
Hmm he is not here.
Dmitri Zagidulin
13:07
I was just thinking the same thing!
Manu Sporny
13:11
same
Orie Steele (Transmute)
13:29
unfortunate
Orie Steele (Transmute)
14:43
Base64url strikes again
Manu Sporny
14:50
yerp
tobiaslooker
15:55
Wait so you sign the CID?
Manu Sporny
16:28
I think that's fine, it's effectively signing the data...
Dmitri Zagidulin
16:33
sounds like. (also, don’t forget to raise hand, ya'll)
tobiaslooker
16:42
Yeah which is really a hash
Orie Steele (Transmute)
16:49
my
Orie Steele (Transmute)
16:50
b
Dmitri Zagidulin
16:52
q+ to ask about chunking
Andreas Freund
17:14
an example might help
Jonak, Sumita T.
18:36
Can you zoom
Andreas Freund
21:07
very deep
Orie Steele (Transmute)
21:19
As deep as you can on encryption and IPFS
Orie Steele (Transmute)
21:58
What JWE alg is used
Orie Steele (Transmute)
22:02
Where is that configured?
Orie Steele (Transmute)
23:18
Should have been called did-josh :(
Orie Steele (Transmute)
23:22
Jose*
Orie Steele (Transmute)
26:55
q+ to ask about CRDTs and IPLD
Dmitri Zagidulin
28:58
blockchain means you don’t need commutativity :)
Andreas Freund
29:22
so the generalized timestamp required is provided by the blockchain
Andreas Freund
29:54
problem is when you have delta's before a state is anchored
Manu Sporny
29:57
/me wonders what the throughput of such a system would be.
Orie Steele (Transmute)
30:20
Statechannels ftw
Manu Sporny
30:42
err, what if you have 3 observers? They can all see the statechannel? how do they sync?
Andreas Freund
31:32
yeah the statechannel must be transparent … in order to know if stuff is being censored
Andreas Freund
32:50
you can also use interval tree clocks if you have no other centralized timestamp provider … they will provide eventual consistency
Dmitri Zagidulin
33:40
(the definitions/reference mechanism looks interesting.)
Orie Steele (Transmute)
36:48
q+ to note hubs
Adrian Gropper
43:26
are you sure it’s narrow?
tobiaslooker
43:34
q+
Manu Sporny
43:39
q+ yes, it's narrow -- if its not, we're in trouble
Adrian Gropper
44:15
can a scope be a subset of a standard data model?
Orie Steele (Transmute)
44:26
Can we agree to a formal model for scopes?
Orie Steele (Transmute)
44:36
Like a way to represent them / document them?
Dmitri Zagidulin
44:47
@Orie - that sounds like a great candidate for a PROPOSAL
Orie Steele (Transmute)
44:51
read:vault etc...
Dmitri Zagidulin
45:01
exactly yeah
Adrian Gropper
45:05
q+
Joel Thorstensson
45:38
Thanks for having me on! It’s quite late here so dropping. See you another time :)
Dmitri Zagidulin
45:45
thanks Joel!
Orie Steele (Transmute)
45:49
thanks!
Orie Steele (Transmute)
46:19
I do not way to make a system that supports more than CRUD + timeing
Orie Steele (Transmute)
47:15
https://identity.foundation/secure-data-store/#authorization-model
Dmitri Zagidulin
47:16
Original blog post “We need to rethink OAuth scopes” that brought up the notion of Structured Scopes - https://medium.com/oauth-2/transaction-authorization-or-why-we-need-to-re-think-oauth-scopes-2326e2038948
Orie Steele (Transmute)
47:20
^ we need to review this
Dmitri Zagidulin
47:37
+1 to review
Manu Sporny
47:53
q+
tobiaslooker
48:03
q+
Dmitri Zagidulin
49:28
Further example of Structured Scopes used in OAuth.XYZ (one of the input specs into GNAP): https://oauth.xyz/transactionrequest/ (the Resources section)
Dmitri Zagidulin
49:42
q+ to mention homework/required reading for structured scopes.
Chris Were
50:42
I’d be very cautious over complicating this. This is low level data storage so should focus on CRUD only.
Orie Steele (Transmute)
50:50
agreed
Dave Longley
50:51
we should do what we need for interop and then mention extension points
Dmitri Zagidulin
50:54
+1
Orie Steele (Transmute)
51:03
We need to document this part
Adrian Gropper
51:10
q+
Orie Steele (Transmute)
51:13
We have interop, but its not documented
Dmitri Zagidulin
53:59
(straw poll) PROPOSAL (from Orie): The SDS WG will specify a data model for scopes / will specify and document a minimal set of authorizations.
Orie Steele (Transmute)
54:12
+1
Manu Sporny
54:20
+1
Dave Longley
54:35
+1
Dmitri Zagidulin
54:35
+1
Chris Were
54:39
+1
Andreas Freund
55:07
+1
Michael Shea
55:38
+1
Adrian Gropper
55:47
+0
Dave Longley
55:50
"scopes" should not imply "OAuth2"
tobiaslooker
56:26
+1 given scope in a general sense describing a scope of authorisation but no direct relationship to Oauth2.0
Orie Steele (Transmute)
56:29
Issue for the proposal: https://github.com/decentralized-identity/secure-data-store/issues/107
Adrian Gropper
56:40
q+
Manu Sporny
57:17
q+
Dmitri Zagidulin
57:23
q+ to give example
Orie Steele (Transmute)
58:01
q+ to respond to adrian
tobiaslooker
58:10
Is that a narrow set that can
Adrian Gropper
58:12
q+ to answer
tobiaslooker
58:16
*can’t be extended?
Dave Longley
58:19
i think adrian may be interpreting "narrow set" to mean "that's all you can do" vs. "this is the minimum you need"
tobiaslooker
58:31
Is it an exhaustive set or a base
Dmitri Zagidulin
58:35
+1 - we’d just be specifying a minimum
Andreas Freund
01:00:27
+1 orie
Orie Steele (Transmute)
01:01:56
I think I understand
tobiaslooker
01:01:58
Got to go thanks all, great call
Orie Steele (Transmute)
01:02:13
Adrian wants to a allow to delegate fine grained semantic access to her data
Orie Steele (Transmute)
01:02:19
I think it only applies to hubs
Dave Longley
01:02:24
yes, it only applies to hubs
Orie Steele (Transmute)
01:02:54
Need to drop