Let's KERI on together
The video
The subtitles as text
0:00:00.080,0:00:00.800 paracore
0:00:03.520,0:00:12.480 okay and I will show slides all right so this is the presentation I did at IIW called 'Let's
0:00:12.480,0:00:20.640 KERI on together'. I did the presentation, I did the card at IIW with instructions on
0:00:20.640,0:00:26.880 how to pull the software and get ready so you could run alongside the demo if you so chose.
0:00:27.600,0:00:33.120 No one did, which was a little bit disappointing, but the repo is here, it's the KERI, right, KERIpy
0:00:33.120,0:00:39.680 repo the web of trust it's a lot of the all of the networking and asynchronous
0:00:41.840,0:00:49.840 I/O is based on co-routines in a library Sam Smith created called 'hio' right now
0:00:49.840,0:00:54.000 what I'll be running today requires Python 3.9.7 but we're now up to 3.10.3
0:00:55.200,0:00:58.960 and these are the instructions if you wanted to follow along let me get clone
0:01:00.000,0:01:04.320 the repo itself check out the future 'iiw' branch and then just run a pip install
pip install
0:01:05.120,0:01:12.960 of the requirements file, to get all of the dependencies. So most of actually everything
0:01:12.960,0:01:18.720 that we'll be showing today, starts with the KERI command line interface that we call 'kli'.
0:01:18.720,0:01:26.080 It has commands and sub commands much like, if you're familiar with Kubernetes' "kubectl" command line,
0:01:26.080,0:01:29.840 so that for example if you were doing some something with verifiable credentials you would
0:01:29.840,0:01:36.080 start 'kli vc' and then there would be a sub command following after the 'vc' for example 'issue' or 'revoke'
0:01:38.320,0:01:46.640 And then we do have Docker containers that have all of the code pushed and configured
0:01:46.640,0:01:53.440 ready to run so you can pull one of those and do a 'docker run', running a sub shell
0:01:53.440,0:02:00.960 and you can use the kli to perform tasks so the basic structure for the commands in the kli
0:02:01.840,0:02:07.840 the groups that we have, pardon me, the agent group for running our cloud agent which is also run
0:02:09.200,0:02:16.800 in the distributable desktop application we're building using Py installer and
0:02:16.800,0:02:23.200 Electron to bundle KERIpy behind a web-U/I that web-U/I uses the agent
0:02:24.880,0:02:29.600 if you want to perform some delegation commands there's a couple of sub commands under delegate
0:02:29.600,0:02:34.320 if you want to create a distributed multi-sig identifier you would use the multisig commands
0:02:34.320,0:02:38.560 and then the verifiable credentials I mentioned the wallet commands are simply for listing the
0:02:38.560,0:02:44.560 contents of your wallet and by "wallet" I mean verifiable credential wallet and then
0:02:44.560,0:02:50.320 to start witnesses or watchers you would use the watcher group with sub cmnds like 'start'&'demo'
0:02:51.200,0:02:58.640 Steven: So, Phil, all these ..., well everything on this page, is there documentation as to
0:02:58.640,0:03:06.240 what exactly an 'agent' means like do you have the terminology somewhere? Phil: No, there isn't, (Now there is!, ed.)
0:03:06.240,0:03:13.440 but that would be a good thing to add to a getting started I guess. Steven: yeah absolutely, so I mean could
0:03:13.440,0:03:20.880 certainly create like a vocabulary dictionary and then maybe a one-liner description
0:03:20.880,0:03:28.400 of what each like "run cloud agent" does, what delegation means. Phil: yeah that would be great.
0:03:28.400,0:03:30.720 (Available! See the glossary in the Youtube description)
0:03:30.720,0:03:46.800 Look at command line, if you run just 'help' from the top level, it does list... Steven: can you share the screen?
0:03:47.440,0:03:53.360 Phil: I'm sharing just this spreadsheet, I mean just the presentation on there, let me show that desktop
0:03:56.000,0:03:59.200 there we go, okay can you see the command line now? Steven: yeah
0:03:59.200,0:04:04.480 Phil: OK, so if you run 'kli --help' you get a list of the commands and sub commands and they do have
0:04:05.440,0:04:10.000 one liners not all of them are accurate at this moment but that's something that...
0:04:10.000,0:04:15.840 Steven: doesn't matter, it'll be a placeholder in the worst case scenario.
0:04:16.400,0:04:21.360 okay, so for key management the top level commands the the very first thing you almost always do
0:04:21.360,0:04:27.920 is do an 'init' which initializes both your data store and your encrypted key store, 'incept' for
0:04:27.920,0:04:32.960 creating a single-sig identifier, technically you can create multi-sig identifiers with that as well,
0:04:33.600,0:04:37.120 but they're not all that interesting. Because all the keys are stored in the same key store
0:04:39.440,0:04:44.400 the different key events that you can create an interaction event or a rotation event
0:04:45.040,0:04:48.480 are handled with the next two commands 'interact' and 'rotate' and they each take
0:04:48.480,0:04:54.160 parameters, 'rotation' for example, specifying your signing thresholds number of keys rotating in
0:04:54.160,0:05:00.000 and out witnesses all the things that you can do when rotating and 'purge' just deletes your key store
0:05:01.600,0:05:09.040 so then the miscellaneous commands querying for creating witnesses for a KEL. Sending your
0:05:09.040,0:05:14.240 KEL to another agent; we don't use that very much anymore because we use OOBIs to accomplish that now.
0:05:15.120,0:05:23.120 Signing, pardon me, you can just sign our arbitrary data and it will return the signatures based on
0:05:23.120,0:05:27.520 the identifier that you used and we'll be done I'll be demoing that in a little bit. Verifying
0:05:27.520,0:05:32.000 those signatures and if you're you know if you're just exchanging signed data with someone via an
0:05:32.000,0:05:36.800 e-mail this will accomplish that for you. Then 'list'ing the identifiers that
0:05:36.800,0:05:42.880 you're creating and giving a 'status' of a single identifier. So it's important to point out that
0:05:44.240,0:05:51.840 with the most recent updates to KERIpy, we now have first class support (ringing silenced)
0:05:57.040,0:06:02.480 We now have first class support for multiple identifiers multiple local identifiers,
0:06:03.760,0:06:07.040 so that's why we have the 'list' command so you can create in you can run
0:06:07.040,0:06:10.560 'init' once create a single data store and key store and then run 'incept' as
0:06:10.560,0:06:13.600 many times you want to create other identifiers and you use those for
0:06:13.600,0:06:20.320 like peer-to-peer communications and things like that, so that you can't be correlated. All right
0:06:22.640,0:06:29.120 so each of the three top level runnable components an agent a witness or a watcher each have
0:06:31.280,0:06:39.520 various startup commands. You have the 'bootstrap' command to start an empty agent or to run a single
0:06:39.520,0:06:44.800 agent against one that's already been started. You can run 'start' and then both for witnesses and for
0:06:44.800,0:06:50.480 agents we have these 'demo' commands which we make heavy use of for local testing. So, for example
0:06:50.480,0:06:55.920 under the witness this starts up three well-known witnesses and by "well-known" I mean we use salts
0:06:55.920,0:07:02.720 to initialize their key stores so that the you can predict what identifiers will be created and
0:07:02.720,0:07:09.760 again that's not anything that you'd want to do in a production environment,
0:07:09.760,0:07:14.240 but for running tests, and we'll be using particularly the 'demo' witness command later on
0:07:15.120,0:07:19.200 to start witnesses so that other commands that need witnesses know where to find them.
0:07:20.320,0:07:26.720 And then we have similar commands for watchers. Part of the demo today is running through a
0:07:26.720,0:07:31.920 series of scripts that we've created in the demo directory located in keripy/scripts/demo
0:07:33.360,0:07:38.960 and the most basic one is just 'demo-script' which starts off and runs all the really basic commands
0:07:39.520,0:07:43.840 'demo-witness-script' runs through some of the same commands but requiring witnesses to be running
0:07:44.720,0:07:50.160 and then we get into the more interesting stuff: delegation and multi-sig
0:07:50.160,0:07:54.800 multi-sig delegation where both parties of the delegation are a multi-sig identifier
0:07:54.800,0:07:59.920 and then credential issuance and multi-issuance. During the demo at IIW we did not get to these.
0:08:00.720,0:08:05.360 We only got this far and in an hour. I'm not sure we'll even get to these, but we will see.
0:08:08.320,0:08:15.760 Oh, I forgot to point out the all of the sample JSON files are also located in keripy/tests/app/cli,
0:08:15.760,0:08:25.360 which contains the ..., when you do a 'kli incept' command, it takes a file for all the properties
0:08:25.360,0:08:30.560 that you want to pass in for the creation of that identifier, for example how many signature or how
0:08:30.560,0:08:34.880 many public keys to create, what's your signing thresholds, how many / what witnesses to use, whether
0:08:34.880,0:08:39.840 it's transferable or not, et cetera and I'll show these as we go through them today
0:08:42.320,0:08:47.360 and then finally another big enhancement in this most recent round of changes to KERIpy
0:08:47.920,0:08:53.840 was the addition of out-of-band introductions or OOBIs. They are files,
0:08:54.560,0:09:00.400 these are startup files that contain bootstrap OOBIs. So for example when we use the 'demo'
0:09:00.400,0:09:04.800 command for witnesses, like I said: they're well-known identifiers, they also start up
0:09:04.800,0:09:08.720 on well-known ports and they publish those ports via a series of OOBIs. This is a
0:09:08.720,0:09:14.880 configuration file that any other agent can run and know exactly how to contact those witnesses
0:09:16.000,0:09:19.120 and this is the configuration file for those witnesses, so we'll get to those in a minute.
0:09:20.000,0:09:26.080 So that's it for the, pardon me, that's it for the presentation there's not slides anyway
0:09:26.080,0:09:30.240 there's not that much in terms of slides, here's just enough to get this kick-started.
0:09:32.720,0:09:38.560 So now we'll go over to our first script today. Oh, one thing I didn't point out, but I wanted to
0:09:38.560,0:09:44.000 make clear: because we have the kli which you can do everything you need to do with KERI
0:09:44.640,0:09:48.640 in KERIpy with the kli on a command line and that's for running in a mode where you're
0:09:48.640,0:09:54.640 not always on the internet. But as I pointed out also, we have an agent that can run.
0:09:54.640,0:09:58.800 That could be a continuous or persistent connection to the internet. So a cloud agent for example.
0:10:00.240,0:10:05.760 And in that case the cloud agent exposes a series of REST APIs that perform the exact same
0:10:05.760,0:10:11.680 functionality as all the kli things that we'll be going through today. And in the demo scripts
0:10:11.680,0:10:17.440 that you see here for every script starting with 'delegate', there is a sister script next to it,
0:10:17.440,0:10:22.320 '-agent', that is the exact same command but with a series of CURL POSTs, I'm sorry, the
0:10:22.320,0:10:28.560 exact same functionality, but executed as a series of CURL POSTs against an agent, or PUTs
0:10:31.600,0:10:35.440 (and I'll show when we launch) the agent we also expose a swagger U/I,
0:10:35.440,0:10:42.000 which allows you, if you wanted, to execute commands against it. I'll show that towards the end.
0:10:44.880,0:10:47.280 So let's start with the basic 'demo' script.
0:10:49.840,0:10:55.920 I'll zoom in here a little bit, is that legible for you? Steven: yeah, I'm much better now.
0:10:57.760,0:11:05.760 Okay so, the very first thing this does is create a key store within it. Currently KERI now supports
0:11:06.560,0:11:11.360 the ability to create a key store and data store for that matter with a passcode and when you
0:11:11.360,0:11:16.400 do that, it becomes an encryption key. like it's stretched into an encryption key that is used for
0:11:17.440,0:11:24.160 encrypting the key store so that no one else can gain access to your private keys
0:11:24.800,0:11:29.040 that passcode is never stored anywhere within KERIpy or the agent
0:11:29.040,0:11:33.520 so it has to be protected by the user because otherwise you can't get access to your keys
0:11:34.400,0:11:41.840 Steven: Sorry Phil give me one sec there's just something going on in here, one second
0:11:45.000,0:11:55.000 (What a perfect break for a small quiz!:) (How many new terms has Phil introduced so far?)
0:11:55.500,0:12:05.520 (The answer is: 55 terms that need explanation!) (Most of which are available; see description.)
0:12:07.520,0:12:08.960 Steven: sorry about that! Phil: no worries
0:12:15.840,0:12:20.240 Okay so, for all of the commands that we're gonna be running today, we just pass in the
0:12:20.240,0:12:23.200 flag '--nopasscode' so the keystore isn't encrypted. And that's just to make things
0:12:23.200,0:12:27.120 easier, so we're not passing passcodes around to every other command but anywhere where you
0:12:27.680,0:12:32.640 run a subsequent command after 'init' you can pass in '--passcode' and it'll unlock
0:12:32.640,0:12:38.400 the keystore if it's indeed locked so the very first thing that this does is create a database
0:12:38.400,0:12:43.680 and key storage you can see from the comments the name of the database, the name of this
0:12:44.480,0:12:48.240 instance that you're creating, is called 'test' and that is used to create
0:12:48.240,0:12:53.440 the directory structure for the databases that support both the wallet and the key store.
0:12:55.520,0:13:00.800 Those databases can go in one of three places. When configured properly you can
0:13:00.800,0:13:06.640 create a database and key store in '/temp' which obviously is used just for testing and
0:13:06.640,0:13:13.360 we use that heavily when running all of our unit tests. If you have a directory called
0:13:14.080,0:13:18.720 '/user/local/var/keri', and the current user has write access to that directory it'll create
0:13:18.720,0:13:22.480 all the databases under there. And then they're prefixed with the name that you give it here.
0:13:23.680,0:13:28.720 If you don't have that directory the last place that KERIpy attempts to create the data stores
0:13:28.720,0:13:34.960 is in your 'home directory.keri' and then it would begin with 'test' for this one after that.
0:13:35.600,0:13:38.160 So those are the three places you want to go look for your key stores
0:13:41.280,0:13:44.240 So after it creates the key store, we're creating... Steven: Sorry, can you, when you
0:13:45.760,0:13:51.280 would you define which one of those three places to look for here? Phil: No. Steven: or it's just the way
0:13:51.280,0:13:56.320 you're installed, you're just, you're set up Phil: that's correct, it's the way your system is set up
0:13:56.320,0:14:01.760 there. we don't support the creation of temporary from the command line. We could add it as a flag,
0:14:01.760,0:14:06.400 I guess, but we don't currently support that. So this will ...,from the command line, this will
0:14:06.400,0:14:09.920 create the key stored in either '/user/local/var/keri', if it exists and you have
0:14:09.920,0:14:16.720 write access to it, or under '.keri' in your own directory if you don't have that other one.
0:14:20.080,0:14:20.580 Uhm...
0:14:23.600,0:14:28.240 Yeah, we could we could definitely add the ability to create a 'temp'(orary ed.) from here but that would
0:14:28.240,0:14:32.960 just be another flag so the first thing we create is a non-transferable identifier after you give
0:14:32.960,0:14:37.520 the name of the data stored you give a local human readable alias to this identifier and as
0:14:37.520,0:14:41.360 i mentioned you can support you can create multiple identifiers so you would give them
0:14:41.360,0:14:43.920 per a single key store so you can give them whatever aliases you want
0:14:43.920,0:14:47.600 So that you can remember what you're using that particular identifier for
0:14:48.880,0:14:52.480 So this is creating a non-transferable identifier. We'll start by opening that
0:14:56.560,0:15:04.320 So, 'transferable=false' obviously and a transferable identifier (<- slip of the tongue!)
0:15:04.320,0:15:08.160 Or a NON-transferable identifier in this case, means that you cannot perform
0:15:08.160,0:15:12.320 any rotations against this (identifier ed.) You can't transfer to another set of keys that's why it's called
0:15:12.320,0:15:19.280 "non-transferable", so these are basically throwaway identifiers. You use them for example witnesses
0:15:19.280,0:15:23.200 you'd use non-transferable identifiers because they themselves won't ever have any witnesses
0:15:23.200,0:15:27.040 they don't perform rotations at all. If you feel like a witness
0:15:27.040,0:15:30.640 has been compromised, you just destroy that witness and start over with a brand new one,
0:15:30.640,0:15:34.720 and rotate it out of your key store, out of your identifier, that's using that witness.
0:15:37.120,0:15:42.560 So because it has no witnesses, 'toad' is the witness threshold I have no idea why it's
0:15:42.560,0:15:48.800 called "toad"; that's a Sam-ism (Sam Smith, ed.). so we have a zero witness threshold, because we have no witnesses the
0:15:48.800,0:15:54.320 'icount' is the number of signing keys to create, the 'ncount' is the number of next keys. If you
0:15:54.320,0:15:59.440 you're familiar with KERI pre-rotation, this is the next keys to create and that these are the
0:15:59.440,0:16:05.960 signing thresholds, 'i' for current signing threshold ('isith' ed.) 'n' for next signing threshold ('nsith' ed.), and these have to be
0:16:05.960,0:16:10.360 '1' for transferable identifiers, across the board, because they only support single sig
0:16:12.400,0:16:18.240 All right, so, after we do that, we try and rotate this identifier and we capture the
0:16:18.240,0:16:23.200 fact that we should get an error, because you can't rotate a non-transferable identifier.
0:16:25.360,0:16:36.080 Then we move on and we create a transferable identifier. Open that file
0:16:36.080,0:16:38.320 the only difference between this and the previous one, because we're not
0:16:38.320,0:16:42.160 going to do witnesses quite yet, is that we have the flag 'transferable:' equals 'true'
0:16:44.000,0:16:46.800 and you'll notice when we create those two identifiers they have different
0:16:47.520,0:16:53.520 prefixes on the identifier which because all cryptographic material is encoded in
0:16:53.520,0:16:57.920 qualified Base64 we have a derivation code on the beginning that indicates one is a
0:16:57.920,0:17:03.840 non-transferable, and the other is a 'Blake3' hash of a transferable identifier; it's 'B' versus 'E'.
0:17:11.120,0:17:14.720 So after we create the transferable identifier we then try and perform a
0:17:14.720,0:17:19.520 rotation event, which should work because this is transferable. We then rotate, and we anchor
0:17:19.520,0:17:22.480 data into it (using '--data' tag, ed.), and in this case the anchor is
0:17:24.560,0:17:29.520 just a simple standard anchor for example for performing a "delegation"
0:17:29.520,0:17:36.400 or anchoring a transaction event log (TEL ed.) event into this, and by anchoring we mean making
0:17:36.400,0:17:40.960 a cryptographic commitment to whatever this data points to. So, for example in delegation,
0:17:40.960,0:17:45.840 if I were being a delegator and someone requested a rotation event, I would anchor,
0:17:46.800,0:17:50.640 I would create an anchor similar to this this is okay this wouldn't be '0' but
0:17:50.640,0:17:57.000 okay, this is the prefix (<- Oops!) that I'm approving, the AID, sorry, that I'm approving
0:17:57.280,0:18:01.360 this is the sequence number of the event that I'm approving and this is a digest of that event
0:18:02.640,0:18:05.040 same thing if you were anchoring an issuance event
0:18:05.840,0:18:09.440 from a public transaction event log (PTEL ed.), it would look exactly like this
0:18:11.840,0:18:18.240 Steven: And the sequence is for what? Phil: So remember, in key event logs key events are a sequence of events.
0:18:18.800,0:18:23.680 So you do the inception: 's=0', rotation 's=...' etc etc, so that's an index into the
0:18:23.680,0:18:31.520 key event log, that this event represents. We perform an interaction event which anchors
0:18:31.520,0:18:36.320 the exact same data just for testing purposes, obviously. Then we perform a rotation event that
0:18:36.320,0:18:41.120 gives us three keys with assigning threshold of two and then another rotation doing the same
0:18:41.120,0:18:46.000 thing, because remember, this is an interesting thing to remember about KERI,
0:18:46.000,0:18:52.560 because you do pre-rotation you need at least two rotations to get to a new number of next keys
0:18:52.560,0:18:57.680 So when we first created these, if you remember, I had a next key of '1' so after this rotation event
0:18:58.720,0:19:02.240 we will only have one key because that's what we previously committed to but now we
0:19:02.240,0:19:06.000 prepare ourselves to have three keys if we perform another rotation event.
0:19:09.680,0:19:14.080 These are the sign&verify commands that I mentioned earlier. So we sign the data just signing
0:19:14.080,0:19:17.840 again you can have any arbitrary data. You can specify the data on the command line, if you want,
0:19:17.840,0:19:25.200 or if you use the '@' sign here, you can specify a file to sign, and then we take the three signatures
0:19:25.840,0:19:30.160 and pass them into the 'verify' command against the same data, using the same identifier that
0:19:30.160,0:19:34.880 signed them, testing with the same identifier that signed them, and we get whether they're verified or
0:19:34.880,0:19:39.440 verifiable, or not, and then in this last one (command, ed.) I just messed with the signed data to verify
0:19:39.440,0:19:46.320 it, to generate an error on verification. And the last bit here is just testing establishment only
0:19:50.000,0:19:54.000 and that's just another flag that you can pass in when creating an identifier: 'estOnly:'
0:19:54.720,0:20:00.720 is 'true' and that means you can only perform establishment events which are 'incept' or
0:20:00.720,0:20:07.120 'rotate' events and that's a security enhancement to an identifier, so that you can't ever expose
0:20:07.120,0:20:13.280 your public key repeatedly by doing interactions which just sign and anchor data but they never
0:20:13.280,0:20:18.400 rotate keys. And repeated use of a public key weakens it because it exposes it and allows
0:20:18.400,0:20:24.640 people an opportunity to try and work against it and break it. So, with "establishment only" (estOnly tag, ed.),
0:20:24.640,0:20:30.240 every time you do anything, issue a credential, approve a delegation, etc, you're getting a new set of
0:20:30.240,0:20:34.720 "next", you're getting a new set of keys so the keys are only exposed once, when they're first created.
0:20:37.280,0:20:37.780 okay
0:20:40.160,0:20:45.040 All right, so let's go run that, how are we doing here for visibility? Let me zoom in a little bit.
0:20:45.040,0:20:49.840 Steven: a little bit bigger will be good. Phil: How's that? Okay yeah, all right, let me give us a little space
0:20:55.440,0:21:03.840 All right, so the first thing we'll do is, so, the very first thing we always do is delete our
0:21:03.840,0:21:10.560 keystore, because I'm developing KERIpy and working with it every day, I repeatedly am
0:21:10.560,0:21:15.840 deleting my keystore, so if you're going to ever do anything with KERI you get used to running this
0:21:15.840,0:21:20.000 command, because you're creating new key stores for testing, and you don't want to keep the old data
0:21:20.000,0:21:25.600 around, so you'll just see me do that repeatedly throughout the demo today.
0:21:27.680,0:21:31.600 So run './scripts/demo/demo-script.sh' Steven: this is a stupid question...
0:21:33.840,0:21:40.720 so when it's the key stores that are stored in that folder '/var/keri'; what format is it in?
0:21:42.000,0:21:50.560 those are lmdb databases. Lmdb? Lmdb, yeah. it's a read-only, or not read only, it's a
0:21:50.560,0:21:54.480 lightning fast key value store supposedly the fastest database on the planet
0:21:55.120,0:21:57.840 according to Sam (Smith, ed.) according to them as well; that's what they claim.
0:22:00.400,0:22:04.720 All right, so we'll just quickly go through the different commands that I
0:22:04.720,0:22:08.560 already talked through in the script. We created the key store in the database, and we created a
0:22:08.560,0:22:13.440 credential store as well, and you can see I have write access to '/user/local/var/keri'
0:22:13.440,0:22:19.280 So under keystore (ks, ed.) created test under database (db, ed.) test and 'reg' for the registry for the key
0:22:19.280,0:22:24.720 credential store, it created test. This is the non-transferable identifier that we created. It has
0:22:24.720,0:22:29.760 a derivation code of 'B' at the beginning of it and interestingly enough the public key is also
0:22:30.880,0:22:37.120 the identifier. Here's the error that we generated because we attempted to rotate a non-transferable
0:22:37.120,0:22:42.000 identifier. That's a good error and then next we created the transferable identifier
0:22:43.440,0:22:47.280 and it starts with 'E' so you know that's a transferable identifier just by looking at
0:22:47.280,0:22:52.320 it. And the public key is different because this public key, I'm sorry, this identifier (!) is generated
0:22:53.520,0:22:59.680 as a hash of the inception event, so that's how we generate identifiers for the
0:22:59.680,0:23:06.560 transferable identifiers and then using the same one, we did the first rotation, we did a second
0:23:06.560,0:23:10.720 rotation, we anchored data in here somewhere, which I'll take a look at with the status command next.
0:23:11.680,0:23:15.120 And then we started doing that we did: an interaction event. And then we did the
0:23:15.680,0:23:23.120 rotation, to get three sets of keys. Okay, that was the second rotation event with the three on it.
0:23:23.760,0:23:28.720 And then here are the three signatures on that anchor data, and those are the same
0:23:28.720,0:23:32.320 that are in the script because I when setting up the script, I copy and paste them in there
0:23:32.320,0:23:36.880 one two and three is valid but signature three is not valid now. Interestingly enough,
0:23:37.520,0:23:42.000 you see that with only passing in the signature the code knew that they were one two and three
0:23:42.640,0:23:46.480 and that's because when using a multi-sig identifier when you sign something with it
0:23:46.480,0:23:51.280 you create what's called indexed signatures so the derivation code contains not only
0:23:52.640,0:23:59.040 a code to tell you that it is a signature but also it is an index signature so that you can see 'AA',
0:23:59.040,0:24:07.040 'AB' and 'AC', that's hex for '0', '1' and '2'. So we know that they came from, they correspond to the first public
0:24:07.040,0:24:11.920 key versus the second public key versus the third public key. Steven: Cool, so how come it says signature
0:24:11.920,0:24:17.680 three is valid and then error signature is invalid? Phil: Because this is the one in the shell script that I
0:24:17.680,0:24:25.440 let's see script "demo-script", I ran this one with a messed up signature, so I took signature three
0:24:25.440,0:24:31.840 and I changed the end of it, so it's an example, just to show it that way.
0:24:32.480,0:24:37.600 i should change the anchored data, because changing signatures is not as interesting. But anyway,
0:24:37.600,0:24:41.360 one of the commands that I talked about at the beginning that's very useful is the 'kli
0:24:42.400,0:24:53.680 status' command and we'll take a look at the test key store and the alias of trans and you can see
0:24:53.680,0:24:57.680 that it's at sequence number five, because we performed an interaction
0:24:57.680,0:25:01.600 in several rotation events. We're not using any witnesses so this is blank, and these are our
0:25:01.600,0:25:05.840 three public keys but more interesting with the 'status' command is this the '--verbose' flag
0:25:07.040,0:25:10.320 which gives you the full key event log so now you're getting into the real guts
0:25:10.320,0:25:14.000 of kerry here you can see exactly, oops, that's not what I want to do,
0:25:19.040,0:25:19.760 you can see
0:25:21.840,0:25:25.360 well first we specifically list the witnesses when you run 'verbose', which he
0:25:25.360,0:25:28.480 doesn't have any, but now you can see each of the commands. So here's your inception command,
0:25:29.040,0:25:33.600 because we use self-addressing identifiers (SAID, ed.) for all KERI events, his digest is the same as his
0:25:33.600,0:25:38.160 inception, is the same as his identifier because this is an inception event, and here's his key
0:25:38.160,0:25:41.840 and here's a cryptographic commitment to his next key, his pre-rotating key.
0:25:42.640,0:25:47.280 Now we perform a rotation so you could you could run Blake3 against this and prove that
0:25:47.280,0:25:51.440 it matched up with the pre-rotated next key, to verify that this is the correct rotation
0:25:51.440,0:25:56.800 event and his identifier of course doesn't change. That's the whole power of KERI for
0:25:56.800,0:26:01.760 key management. But he has a new self-addressing identifier, because this is the hash of this event.
0:26:03.680,0:26:08.400 Then we move on and this is where we anchor the data, and this is what an anchor looks like.
0:26:08.400,0:26:13.920 So it's simply in the 'a:' field in a which is an array of any data that you want to anchor in here.
0:26:13.920,0:26:19.040 And this is that anchor file that I showed you earlier. Here's an interaction event. You see
0:26:19.040,0:26:22.960 there's nothing specifying any key changes here, because you don't change keys with interaction
0:26:22.960,0:26:27.200 events, you're just at anchoring data and here we've anchored the same amount of data.
0:26:28.480,0:26:33.120 You notice also every event but the inception event has a 'p:'
0:26:33.120,0:26:39.680 and that's an anchor to the previous; thus making it a blockchain. An anchor to the previous event
0:26:41.600,0:26:44.960 and then finally we performed the rotation,
0:26:46.080,0:26:50.640 to give us three next keys, but as I said, because we had already committed to one key when we did
0:26:50.640,0:26:54.800 that three, we got three next keys but only one public key, so we had to perform a second
0:26:54.800,0:27:01.360 rotation to get three public keys, and this 'kt:' here is the signing threshold for the (signing, ed.), so you
0:27:01.360,0:27:05.840 need at least two signatures to make anything valid for this identifier in this current state.
0:27:08.000,0:27:11.840 All right. So that's the first intro script. Any questions about that?
0:27:13.040,0:27:14.480 Steven: I don't know enough to ask questions.
0:27:18.560,0:27:24.400 So I know from reading some of the specs of that all the those (tags, ed.) like 'vt:' whatever the
0:27:24.400,0:27:29.760 field names are, are all defined in the spec, right? Phil: yes, they are, absolutely. Yeah, it's funny,
0:27:29.760,0:27:35.120 I mean, I'm at the point now, where I look at them and my head just reads them out loud to me. Steven: I'm not
0:27:35.120,0:27:44.640 there yet. Phil: yeah, it takes a while and Sam has changed them repeatedly. Like 'k' here used or i
0:27:44.640,0:27:48.240 know b all the b's they used to be 'w's because they used to be witnesses but we changed them
0:27:48.240,0:27:54.160 to backers instead of witnesses so now they're all 'bt', 'br' and 'ba' Steven: But is a backer a witness?
0:27:55.840,0:27:59.760 Phil: Well, you can have either a ledger backer or a witness backer that's why we changed the name.
0:28:01.360,0:28:05.680 and a witness backer would be using standard KERI infrastructure for witnesses
0:28:05.680,0:28:09.360 and then you can have as many as you want, but if you have a ledger backer you're anchored to just
0:28:09.360,0:28:15.840 one ledger. And that ledger serves the role of both witnesses and technically watchers.
0:28:16.400,0:28:20.800 Steven: a ledger is greater than ... a ledger contains witnesses?
0:28:22.320,0:28:28.880 Phil: a ledger serves the role of a witness, if you have a need for distributed ledger technology in your
0:28:28.880,0:28:33.920 infrastructure. iIf you're relying on cryptocurrency or you know the economics of a token or whatever,
0:28:34.960,0:28:42.240 then you can use that and trust the (if you want to put trust in the) anchoring of your
0:28:42.240,0:28:47.440 events in the ledger, instead of getting receipts from witnesses, you can do that.
0:28:47.440,0:28:52.720 It's not better, it's just different. It's just a different way to provide that layer of security.
0:28:54.960,0:28:58.160 So the next thing I'm going to do here, down in the lower window, is to launch..
0:28:58.160,0:29:01.280 Steven: So, it has a story (as a sidebar): Has there been discussion about using tokens?
0:29:02.640,0:29:08.000 Steven: Not for us (KERI/ACDC, ed.), no. Steven: But like crypto? Phil: At IIW last week there was a really interesting session
0:29:08.560,0:29:15.680 that Richard Esplin from Evernym (now AVAS) held, asking why KERI versus why a ledger?
0:29:15.680,0:29:20.560 And what the whole point of it for him was... (<- cliff hanger!), that it was a great session, he held it on Monday and none of
0:29:20.560,0:29:25.840 us were able to attend. So then he re-held it again on Tuesday, oh I'm sorry it was on Tuesday, first day
0:29:26.400,0:29:32.560 on Tuesday, then he held it again on Wednesday after talking to us Tuesday night at dinner he
0:29:32.560,0:29:38.320 held it again on Wednesday, so Sam and myself and Kevin could participate, pardon me, and his
0:29:38.320,0:29:42.240 the whole point of it for him was ( tadaa! -> ) okay, if I'm instructing clients who want to build
0:29:42.240,0:29:46.400 SSI infrastructure, do I tell him "he is KERI" do I tell him "he's a ledger" what do I tell him?
0:29:48.160,0:29:51.280 And it was great to clear up a lot of misconceptions because people think:
0:29:52.000,0:29:56.400 "oh, you can't use a ledger with KERI". No that's not true, you don't have to use a ledger with KERI. We
0:29:56.400,0:30:01.200 don't think it's, we don't think it's needed at all, but as Sam as what Sam pointed out during
0:30:01.200,0:30:05.920 the talk, is if you already have an investment in ledger infrastructure for other reasons,
0:30:06.640,0:30:10.720 you can then leverage that ledger as an additional trust threshold,
0:30:10.720,0:30:14.640 or security threshold for your identifiers instead of using witnesses.
0:30:17.200,0:30:22.800 So no, we're not, I mean the vLEI won't be using anything to do, with cryptocurrencies or tokens but
0:30:22.800,0:30:29.280 if you, you know, if for example a lot of people are using SSI systems based on Indy networks, right,
0:30:29.280,0:30:34.320 now if you want to transition to using KERI, you could do that by anchoring your KERI identifiers
0:30:34.320,0:30:39.040 in your Indy ledger. And then you, and the neat thing about it is, you could then transfer them
0:30:39.040,0:30:43.600 off, off the ledger and have non-ledger based identifiers, so that's what's great about
0:30:43.600,0:30:47.760 KERI: it's portable, you can be anchored to any one ledger at a time, or you could move it to a
0:30:47.760,0:30:53.040 different ledger, or you could move to using just witnesses, all with the same identifier by just
0:30:53.040,0:30:58.880 doing rotation events and changing your anchor, your backers here. Steven: but could you also do, if
0:31:02.080,0:31:09.520 could you be anchored, let's say to multiple Indies or Ethereum? Phil: you could be only anchored in one at a
0:31:09.520,0:31:16.160 time, because that's the source of truth for that identifier at that given point in time.
0:31:16.160,0:31:21.440 You can then rotate to a different one, but you you can't have two ledgers at the same time.
0:31:23.200,0:31:30.320 Steven: So when people, well what maybe they misunderstood, so I understood that
0:31:30.320,0:31:33.360 okay one of the problems in blockchain is that you're all you have to be committed to all use
0:31:33.360,0:31:39.120 the same network which is not practical unless you're running an ICO scam, so
0:31:40.960,0:31:46.880 I understand KERI is digital ledger technology agnostic. And that you could
0:31:46.880,0:31:52.080 use it..., it could interoperate with existing ledgers like Aries or Indy whatever it's called,
0:31:54.800,0:31:57.760 but it's only that one network. You couldn't, you couldn't connect
0:31:59.280,0:32:07.520 Ethereum and Indy to the KERI ledger (<- KEL?!) Phil: Correct, yeah, KERI is not (a ledger. ed.), KERI couldn't,
0:32:07.520,0:32:13.520 It couldn't be used as a technology to bridge between networks, that's not what it's designed for, but
0:32:13.520,0:32:17.200 you could use it to move a single identifier across networks, if you wanted
0:32:19.360,0:32:29.520 oh okay, all right, yep, okay. So moving on, the next command I'm running again
0:32:29.520,0:32:34.160 we do the 'rm -rf' and then 'kli witness demo'. This starts these three witnesses and
0:32:34.160,0:32:38.320 we give them aliases of 'wan', 'wil' and 'wes' and these are the three identifiers.
0:32:38.320,0:32:41.840 That I now know by heart, because I start them a hundred times every day.
0:32:42.800,0:32:48.240 And they're started with a given salt for each one, so that the identifier is predictable.
0:32:50.000,0:32:54.640 Now we get to OOBIs. This is the really interesting part. So OOBIs or "out-of-band introductions"
0:32:54.640,0:33:00.720 allow people to discover full key event logs, witness key event logs and service endpoints.
0:33:00.720,0:33:06.800 for any other identifier, and you can see here they are simply URLs, so in these cases this is
0:33:06.800,0:33:13.760 an OOBI, we know that because it has this '***/oobi/' in the path. This is the identifier about whom
0:33:13.760,0:33:18.960 this OOBI is referencing. And this is the role that this particular OOBI is playing, so in this case
0:33:18.960,0:33:26.880 this is a 'controller' OOBI for this witness. We'll take a look at two configuration files.
0:33:33.200,0:33:38.320 So this is a new (setup? ed.), with the most recent version of KERI with Py, we have the ability
0:33:38.320,0:33:42.160 to specify configuration files. Previously when we did our first pilot for life with
0:33:42.160,0:33:44.640 this we were hard-coding IP-addresses all over the place, because we didn't
0:33:44.640,0:33:48.560 have a discovery mechanism. But we now have a discovery mechanism that's using OOBIs.
0:33:49.680,0:33:54.000 So when you start any agent who's going to run with persistent connection like our witnesses,
0:33:54.000,0:34:00.320 or if you run agents, you can specify a CURLs field in their configuration file,
0:34:00.320,0:34:06.880 that specifies endpoints that he will be exposing. So in this case the scheme he's exposing a TCP
0:34:06.880,0:34:15.600 end-point. On this port 'localhost' and an 'http:' endpoint on this port on localhost when he generates OOBIs
0:34:15.600,0:34:19.440 he can generate OOBIs, for either of these endpoints, because he's been told, through this
0:34:19.440,0:34:26.320 configuration file that that's who he's listening on. So when he generates an OOBI for example for a
0:34:26.320,0:34:30.720 controller, so then the controller is basically saying: "this is where you can contact me directly, not through
0:34:30.720,0:34:38.480 a witness, but just contact me directly", pardon me, he would generate a newbie that looks like this...
0:34:41.520,0:34:45.280 Okay, we're already schooled over this, is the little window of the two of us was blocking it
0:34:47.600,0:34:52.560 so it's again, it's the location to contact this OOBI which is of course the same
0:34:52.560,0:34:57.760 as his 'http:' endpoint: what they identify, who the identifier is, and what role this OOBI is playing.
0:34:57.760,0:35:02.480 Now interestingly, when you look at the command line we ran six OOBI resolutions, the
0:35:02.480,0:35:07.440 little green checkboxes, said they all worked and that's because, well now when we start this
0:35:08.000,0:35:12.560 based on the configuration file, not only does each witness expose his endpoints this way,
0:35:12.560,0:35:19.360 but he also resolves the other two witnesses OOBIs, so at startup he exposes this then he
0:35:19.360,0:35:27.280 reaches out to this witness at this address and this witness returns a reply event
0:35:27.840,0:35:32.320 which contains his key event log, which is very simple, because he's a non-transferable
0:35:32.320,0:35:38.080 identifier, and a reply message containing his endpoint information that he has had
0:35:38.080,0:35:41.920 configured via this configuration file. So now all the witnesses are communicating with each
0:35:41.920,0:35:45.360 other and establishing how they could connect with each other if they need to. And also if
0:35:45.360,0:35:50.160 someone is, let's say, as we will demo in a minute, someone is using all three of these witnesses,
0:35:52.080,0:35:56.720 when someone asks a witness for an OOBI for an identifier, that he's a witness for,
0:35:56.720,0:36:01.520 he can say "oh well, he's also has these other two witnesses and here's
0:36:01.520,0:36:05.200 their connection information. Doesn't have to do that, but he can, because we've configured him to
0:36:05.200,0:36:11.120 resolve the other witnesses and then wil's and wes's configuration file is the same.
0:36:11.120,0:36:17.280 Except of course he's referencing the other two identifiers, right, so now we'll look at 'demo-script-witness'
0:36:20.800,0:36:28.320 again, this is we initially initialized the key store with, no passcode, we do an OOBI
0:36:28.320,0:36:32.720 resolution against each of the three witnesses, those OOBIs should look familiar because the ones
0:36:32.720,0:36:38.240 I just showed in the configuration file, so now this new data store now has the information for
0:36:38.240,0:36:42.960 each of these three witnesses. So he's configured to be able to use them as witnesses, if he wants
0:36:43.840,0:36:49.680 this, however, if we were running an agent we would probably create a configuration file like this one.
0:36:50.960,0:36:54.960 If for example we were, which we will be for the vLEI system,
0:36:54.960,0:36:59.600 packaging up an agent that has pre-configured witnesses; that we're anticipating everyone will
0:36:59.600,0:37:03.600 use and then we create a configuration file which gives them OOBIs to those witnesses.
0:37:03.600,0:37:06.960 So when they start up, they initially automatically initialize those,
0:37:06.960,0:37:10.720 and bring them, bring all the data over they need, so they can communicate with those witnesses.
0:37:12.800,0:37:17.120 But you can do it with command line here 'kli oobi resolve' so that's what we do for this.
0:37:19.280,0:37:26.160 So then we do an inception event with the 'trans-wits-sample' and then we create an inquisitor
0:37:26.160,0:37:31.840 and then we do a status of this with his witnesses, we perform a rotation witness
0:37:31.840,0:37:35.520 cut, so that removes this guy from his set of witnesses, this is rotating a witness
0:37:35.520,0:37:39.360 out, because I don't want to use them anymore and then we do a status again
0:37:39.360,0:37:43.840 to see that he's gone, that he's down to two witnesses this is fixed, I fixed this,
0:37:46.080,0:37:51.600 and then we do a kli rotate with a 'witness-add', so we add that witness back in and then we do
0:37:51.600,0:37:56.240 another status to see that he has those, that witness's back in, as part of his witness set.
0:37:56.240,0:37:59.760 Steven: And what happens if you never added that back in? Does something tell you have,
0:38:00.560,0:38:03.040 you should have three witnesses? Phil: Well you're okay,
0:38:04.640,0:38:08.320 you don't have to. It's just saying that you now have two and when you, so when you
0:38:08.320,0:38:18.080 rotate this out, it will also change your witness threshold. I think let's take a look at this guy
0:38:18.080,0:38:22.400 yeah so you can see when we create him he's created his transferable with these three
0:38:22.400,0:38:26.880 witnesses and he has a witness threshold of two so that means you only need receipts from any
0:38:26.880,0:38:31.280 two of these okay considering Steven: Okay, that's sort of, that's what I was thinking / wondering about.
0:38:31.280,0:38:36.720 So if you put that to three and you and you removed, you cut that one...,
0:38:37.440,0:38:42.480 Phil: you'd get an error. Steven: Oh, you get an error, yeah, you can't cut them? Phil: you
0:38:42.480,0:38:46.480 can't, yeah, I would say your witness threshold of two is you're with a threshold of three can't be
0:38:46.480,0:38:50.720 satisfied, because you only have two witnesses. Steven: so you're supposed to create a new one, add a new one
0:38:50.720,0:38:54.320 and then... Phil: No, you can change the threshold on the command line, so you do something like
0:38:55.440,0:38:59.840 '--toad 2' and then that would, that would make it valid
0:39:01.680,0:39:04.240 Steven: but wasn't there a reason why it was three in the first place?
0:39:05.760,0:39:09.440 Phil: It's all your own identifier and you're controlling your identifier, so it's up to
0:39:09.440,0:39:14.560 you to determine what security constraints you want for that identifier, so yeah sure, you
0:39:14.560,0:39:17.920 probably created it with three, but if you're just saying "okay, you know what, now I only need two, or
0:39:17.920,0:39:21.760 I'm only going to have two temporarily, because i need to go find a new witness", you can do this.
0:39:21.760,0:39:27.200 Again not the most practical example but I'm just showing how 'cut's and 'add's work, makes sense?
0:39:27.200,0:39:40.340 Steven: yeah. Phil: All right, cool, so let's run this script and we can watch it go through its process.
0:39:40.840,0:39:44.720 (Quick question: Does Phil enjoy being around Wan, Wil and Wes??)
0:39:44.720,0:39:45.220 yes
0:39:48.720,0:39:51.840 All right, creating key store, this is the OOBI resolution so that this
0:39:51.840,0:39:55.280 guy now knows how to contact those OOBIs. And you can see
0:39:55.280,0:39:58.640 he's going through all the different steps. And we'll go step by step through this.
0:40:02.640,0:40:09.520 Page up, so this is where he initialized his, or he resolved his three OOBIs with the 'oobi resolve'
0:40:09.520,0:40:15.600 command and you can see part of the 'incept' command also takes your identifier and propagates it to
0:40:15.600,0:40:20.560 your three witnesses. However, many end number of witnesses waits for receipts from those witnesses,
0:40:20.560,0:40:25.200 then collects all the receipts and passes them out to all the other witnesses. So he propagates the
0:40:25.200,0:40:30.960 receipts from everybody to every witness. So every witness has a fully receded key event log for him
0:40:30.960,0:40:34.800 and that's what waiting for witness receipts is here. And that all happens as part of the
0:40:34.800,0:40:40.320 'incept' command, then we did the 'status' command so this is 'EELF_' as our new identifier
0:40:40.960,0:40:45.280 and you can see that he has a signing threshold of two but he got all three receipts
0:40:50.480,0:40:54.480 and then as we go on, you can see this is where he cut that one witness, so he's now
0:40:54.480,0:41:00.880 down to two, and this is where he added him back in, he's now back up to three, so we'll do a quick
0:41:04.640,0:41:05.840 test
0:41:16.240,0:41:19.840 my fingers know I made a mistake my brain didn't
0:41:22.080,0:41:22.580 yes
0:41:32.240,0:41:33.040 what did I type
0:41:38.960,0:41:41.520 witness test trans wits
0:41:43.600,0:41:49.600 Okay, so here's the key event log for all those things that we just did. This is the inception
0:41:49.600,0:41:53.360 event, and in an inception event you just specify the 'b:' field with the backers you're
0:41:53.360,0:41:58.480 going to use for this and the 'bt' is your backer threshold so that's the 'toad:' of two that we saw.
0:42:00.480,0:42:07.520 Here's the rotation event that "did a 'br'" or backer remove, so this cut that witness from the list
0:42:08.320,0:42:11.680 and then this is the third rotation, of a second rotation event
0:42:12.240,0:42:17.040 that "did a 'ba'", so it added that backer back into the list
0:42:17.040,0:42:22.000 and when we did the '--verbose', you can see that he lists, he's now back up to, those three witnesses
0:42:24.560,0:42:29.120 All right, so that's using witnesses with an identifier, again I don't think that I show
0:42:29.120,0:42:35.360 that yeah I showed this yeah yeah so you just specify your witnesses in the 'wits' configuration,
0:42:35.360,0:42:38.320 in the 'wits' field of the configuration file, when creating an identifier
0:42:38.880,0:42:46.000 and that's in the demo script, when we did the 'incept', we specified that file 'trans-wits' sample,
0:42:48.400,0:42:51.120 and I used to do a query in this. That's why we're creating this guy "inquisitor"
0:42:51.120,0:42:57.840 here, but we don't do anything with it, but we can take a look at the 'kli list'
0:42:59.040,0:43:03.280 and you can see I have two aliases inquisitor and trans-wits, so I have two identifiers in
0:43:03.280,0:43:08.640 this one key store. And that's who they are and then from there, I could do a 'status' to
0:43:08.640,0:43:14.640 see what each one looks like. All right we got 15 minutes left. We're probably only
0:43:14.640,0:43:17.440 going to get through delegation because, I want to show you the agent real quickly.
0:43:21.120,0:43:27.200 Okay, so we'll go over here, we'll start closing some windows, all right, so let's go over delegation
0:43:31.440,0:43:35.120 So with delegation we're going to create two identifiers. One is the
0:43:35.120,0:43:40.960 delegate, and the other is the delegator. So we create both key stores then we run
0:43:40.960,0:43:45.200 an 'incept' command for. The delegator, using this configuration file,
0:43:50.160,0:43:52.560 he should look familiar, this is just like
0:43:52.560,0:43:55.280 the transferable guy we just created; all the same features.
0:43:59.440,0:44:03.600 Oh yes, here's where we're specifying a salt for
0:44:03.600,0:44:07.840 each of these. So that we know what their identifier will be when we generate them.
0:44:09.360,0:44:14.000 And then we do an 'incept' command for him, I just said that, and then the delegate keystore is
0:44:14.000,0:44:20.800 resolving the OOBI for the delegator. Now this OOBI is different. This OOBI is a witness OOBI.
0:44:20.800,0:44:25.760 So because, as you saw in the configuration file, I'm using these three witnesses, he can generate
0:44:25.760,0:44:32.000 an OOBI which tells anyone interested how to contact him through his witness. So it starts
0:44:32.000,0:44:38.400 with OOBI, just like the other, it starts with the AID or the identifier for whom this OOBI is about,
0:44:38.400,0:44:43.200 and then the role. In this case it's "witness" and then it specifies the identifier of the witness.
0:44:45.440,0:44:45.940 All right
0:44:48.480,0:44:49.840 so let's get started...
0:44:52.960,0:44:55.680 Yeah, so I guess I just I have to run this
0:44:58.800,0:45:02.160 kill and restart, because we want to clean data store. Actually, I don't technically
0:45:02.160,0:45:06.320 for this one, I don't have to, but i'd like to do. (then we hear my grandson crying in the background)
0:45:11.040,0:45:15.680 all right oh sorry one thing I forgot to point out is, this command here when you inset
0:45:15.680,0:45:19.040 let's let's open him so we'll look at delegate.delegatee.json
0:45:21.360,0:45:26.320 so the only difference with him from all of the others that we've done before is that he
0:45:26.320,0:45:32.800 specifies a 'delpre:' so he's saying any key event that I do, any key management event that I do
0:45:32.800,0:45:39.440 an 'inception' or 'rotation' needs to be approved via an anchor by this identifier and the anchors are
0:45:39.440,0:45:43.760 what we saw earlier. You do a rotation event or an interaction event, and you put that data in there.
0:45:43.760,0:45:47.280 And by signing that key event and anchoring it to your key event log (KEL, ed.), You're making a
0:45:47.280,0:45:51.360 cryptographic commitment approving this whatever event he's requesting.
0:45:53.280,0:45:57.920 And we'll take a look at that data, as soon as this is done, so he's only using one witness just
0:45:57.920,0:46:05.040 because I'm lazy, so he will then wait because he has a delegator he'll wait to finish
0:46:05.920,0:46:11.280 until a delegator approves him this command is actually reentrant so if you control see it
0:46:12.240,0:46:16.240 go away call your guy say hey I just sent you a delegation event can you approve it he approves
0:46:16.240,0:46:20.240 it you can come back and rerun it and it'll finish the event propagate the receipts of the witnesses
0:46:20.240,0:46:25.360 et cetera but in the meantime we'll just leave him running and we'll go to another window and then
0:46:25.360,0:46:32.400 Steven: How does it delegate, how does the .., how does that person know, that they're waiting on it?
0:46:33.840,0:46:38.080 This aside from you saying "get on there and approve it", is
0:46:38.080,0:46:43.920 there some messaging? Phil: yes so this 'incept' command will actually notice that he has
0:46:43.920,0:46:51.120 a delegator listed in here, and he will send his key event for his inception event
0:46:51.120,0:46:55.840 to the OOBI, that he was up to the endpoint that he found from OOBI resolution here.
0:46:56.800,0:47:02.400 And the delegator or any identifier, when they receive a key event and they process it
0:47:03.360,0:47:06.640 and they're listed as the delegator, it's a trigger that they have
0:47:07.200,0:47:11.680 something to do and that's what this delegate confirmed command is he wakes up checks his
0:47:11.680,0:47:18.000 mailbox and in his mailbox he finds this event notices that he's listed as the delegator.
0:47:18.560,0:47:24.080 And so then he will ask, you'll see on the command line, he'll ask "Hey so-and-so wants
0:47:24.080,0:47:26.880 you to be their delegate, and you can either approve it or deny it".
0:47:29.200,0:47:35.200 I guess "reject" is the better word, so did I do this yet, I'm pretty sure again
0:47:36.720,0:47:39.520 All right so we're going to run this, he's resolving the OOBIs
0:47:41.600,0:47:46.560 for the witnesses, and you can see that he resolved the delegator's OOBI right here.
0:47:46.560,0:47:50.000 And then he says waiting for delegation approval so he's just sitting here,
0:47:50.000,0:47:56.880 waiting scanning for someone , for this person, to approve his event, so I'll just copy and paste this.
0:47:59.840,0:48:05.840 So he wakes up, he discovers the inception event, so he says a delegation inception request,
0:48:07.360,0:48:13.200 a delegation inception request from this identifier, "accept or deny, I will accept it". He
0:48:13.200,0:48:19.680 signs the event, he propagates his own event to his witnesses and then sends the fully receded event
0:48:19.680,0:48:25.600 back to the delegator, but to the delegate who then says "Okay, I am now an approved event
0:48:25.600,0:48:28.800 and I'm going to get my witness receipts, and here is my new identifier"
0:48:30.320,0:48:34.880 We'll look at two different status, so we'll do 'kli status'
0:48:38.080,0:48:39.040 I think it's delegate
0:48:41.760,0:48:46.320 It is. So you can see because he's a delegated identifier. The status command now prints out
0:48:46.320,0:48:51.040 this bit of information: here's my delegator and the fact that he's anchored, and this is
0:48:51.040,0:48:55.360 important, because if I were to control scene and "do a 'status'" I could come back in three
0:48:55.360,0:48:58.720 days later and do a status it doesn't say "shoot, he hasn't been anchored yet". I need to
0:48:58.720,0:49:02.400 get on the phone and bug him or whatever or run the incept command again and see if the anchor
0:49:02.400,0:49:07.680 is out there for me to process and then he's done his witnesses. So if we look at a '--verbose'
0:49:08.720,0:49:13.520 here, this is his inception event because it's a delegated identifier, it's a different tag,
0:49:13.520,0:49:18.560 or sorry, different type for the inception event, a 'dip' instead of an 'icp'
0:49:20.000,0:49:25.520 and the only other difference here is the last field is a 'di' which is the delegating identifier
0:49:26.480,0:49:31.680 and that is the delegator, so then anyone processing this will know that this is a delegated
0:49:31.680,0:49:36.800 identifier and any event that he processes has to be approved by this guy, for it to be a valid event
0:49:37.360,0:49:44.880 Now notice here his identifier is 'EY7' because it's an inception event, I said that backwards,
0:49:44.880,0:49:49.120 (tries again: -> ) This is the identifier, because there's an inception event, the SAID of this is the same,
0:49:49.680,0:49:57.840 so these are the two, and here's a sequence number '0'. So now, if we take a look at the delegator
0:50:02.400,0:50:09.280 you can see his rotation event here, this is the approval and that 'EY7'
0:50:09.280,0:50:15.040 is the identifier as well as the SAID or digest of the event that he approved and
0:50:15.040,0:50:20.640 he has anchored that data; thus approving the delegation request from the other guy.
0:50:24.240,0:50:29.520 All right so that makes sense? Steven: Maybe one day, yeah, that's what I'm feeding,
0:50:31.200,0:50:35.200 it's so inspiring. Phil (laughing): "yes there's a lot, there's a lot here.
0:50:37.040,0:50:41.360 All right we have eight minutes I will try and get through multisig actually you know what no
0:50:41.360,0:50:45.920 Steven: No rushing in my account, I mean, I understand you're on vacation here, but don't worry
0:50:45.920,0:50:51.440 about my time. Phil: Yeah, I know, I have another meeting, I have to get to. Steven: Oh okay, not only
0:50:51.440,0:50:57.120 for visiting the family? Phil: yeah exactly, yeah but not only that, but I'm out here buying a house,
0:50:57.120,0:51:08.240 because we're moving out here, so all that, as well as work, it's been very very busy.
0:51:08.240,0:51:09.920 All right, so let me show you the agent really quickly.
0:51:12.640,0:51:21.840 So here we start actually. I don't need the witnesses but anyway so we will run 'kli'
0:51:25.760,0:51:31.200 So I'm running 'kli agent demo', just like the demo command down here, he starts,
0:51:32.480,0:51:37.040 except in this case, he starts four well-known agents and that's because, well I needed to test
0:51:37.040,0:51:43.440 multi-sig delegation with two participants in the delegator, and two participants in the delegatee.
0:51:44.320,0:51:52.960 delegation coordination I start him with a ... (that's not gonna work) Well that's fine, I
0:51:52.960,0:51:59.840 started him with a configuration file that tells him where these three witnesses are and he starts
0:51:59.840,0:52:05.200 four agents on these well-known ports again with actually he doesn't have well-known salts because
0:52:05.840,0:52:12.720 you do those via the CURL commands so once you start those agents, let's see, here we go,
0:52:14.960,0:52:22.640 you get on each of those ports a Swagger U/I which contains the available API calls.
0:52:22.640,0:52:29.200 For the agent depending on his state. Every agent initially starts off in a locked state.
0:52:29.200,0:52:33.280 so think of a ..., remember when we did the 'kli init' commands and created the data store
0:52:34.880,0:52:38.160 that data store is now encrypted and locked, if you gave it a passcode.
0:52:38.720,0:52:43.200 So in this case we don't have any data stores yet, so you can start an agent with no data stores
0:52:43.200,0:52:51.840 around and that's when you end up in this state. So in order our passcode requirement for creating an
0:52:51.840,0:52:56.800 encrypted key store is a 22 character passcode we give you just a little helper function. This isn't
0:52:56.800,0:53:01.600 really part of KERI, but we give you a little helper function for, (my grandson is really unhappy),
0:53:04.320,0:53:09.520 for just generating random passcodes if you need that, you can also use "OnePass" or whatever to
0:53:09.520,0:53:16.640 create them and then store them in OnePass, so now just like 'kli init' that creates a key store
0:53:16.640,0:53:23.600 we have a POST to '/boot' which will create a key store for you, you call it, let's try it
0:53:23.600,0:53:29.920 out, let's not type into the sample I did that during the demo too, 'Agent0', and spell agent right
0:53:32.320,0:53:35.440 and give it the passcode, that you saved and execute,
0:53:36.720,0:53:44.160 and you can see the 'agent0' and key store created and he resolved a whole bunch
0:53:44.160,0:53:47.920 of OOBIs, that's part of the configuration file, I guess I do have running the ...,
0:53:48.720,0:53:54.800 So, one thing to note about here, is several of the OOBIs are actually data OOBIs. And we can also
0:53:54.800,0:53:59.440 use OOBIs to resolve things like credential schema. So that's a whole separate topic.
0:54:01.520,0:54:06.880 So, now that we've created the key store the next step is to unlock it, and so this is like
0:54:07.600,0:54:12.560 if you ..., so we ran this demo at IIW after doing our demo hour where we demo-ed the U/I
0:54:12.560,0:54:15.680 that we have sitting on top of this to Keep, and so when you get into the Keep,
0:54:15.680,0:54:19.280 you get a nice screen that says "hey, create your passcode", which is that first API call.
0:54:19.280,0:54:23.840 And it uses the top API called to generate random keystore keys for people who want samples
0:54:24.480,0:54:29.040 to use and then we do a PUT against '/boot' for unlocking. And that's the next
0:54:29.040,0:54:33.040 nice block in the U/I that says "now unlock your key store" Every time you come back into the agent,
0:54:33.040,0:54:37.840 because you already have a key store, you just have to unlock, right, because you've already created it
0:54:41.680,0:54:46.240 So we're going to try it out, here we called him 'agent0',
0:54:48.400,0:54:53.600 and there's the passcode that we use, and we're going to execute this and he says "'agent0' is
0:54:53.600,0:54:58.560 now unlocked", and the cool thing about unlocking an agent, is you just reload and you now have a
0:54:58.560,0:55:03.360 whole new set of APIs for doing all the things you would want to do with an agent, and these all
0:55:03.360,0:55:09.520 marry up very nicely with the 'kli' commands that we've been running. Get your list of identifiers,
0:55:11.360,0:55:15.280 get information about a specific identifier, you can create a new identifier, this would be
0:55:15.280,0:55:23.600 like 'kli incept', you can update, so again not part of KERI but to support a user interface, a wallet
0:55:23.600,0:55:26.960 for example, where you're going to have a ton of different identifiers, you need to associate
0:55:26.960,0:55:32.160 metadata with those. You give yourself context to remember, you know what organization, who ...,
0:55:32.160,0:55:36.560 Just think of a standard contact book; that's what this API is for us; for updating that.
0:55:37.600,0:55:42.160 And because we're "zero trust", all that data is signed at REST, so that if you lose control of
0:55:42.160,0:55:47.680 your database, people can't put in like, you know, change the name of the contact and fake you out,
0:55:47.680,0:55:52.960 because we verify the data when we reload it, and then you can do a rotation or an interaction event.
0:55:53.840,0:55:58.400 This is for creating and listing your credential registries so the public transaction event logs
0:55:58.400,0:56:03.520 for anchoring issuance and revocation events a presentation request, you can generate
0:56:03.520,0:56:07.520 your own OOBI or resolve other people's OOBIs. This is like the 'kli oobi' commands
0:56:08.400,0:56:13.360 these are the this is challenge response so one of the requirements of the Keep is that you do a
0:56:13.360,0:56:18.160 two-factor auth, so this is for generating like a list of challenge words, that's what this guy does.
0:56:18.160,0:56:23.040 Which is just randomly generated challenge words. You would then take those, send them to another
0:56:23.040,0:56:29.280 person using this command. They would then get a notification that someone has challenged them.
0:56:29.280,0:56:33.040 They sign it with that identifier and send it back and now you have proven that they have
0:56:33.040,0:56:38.000 control of your identifier, and that was actually a big part of the demo, that we did during demo hour.
0:56:41.360,0:56:45.600 This is more information about getting, about getting an updating contact information.
0:56:45.600,0:56:49.680 We also support images. We don't sign those yet and Samuel (Smith, ed.) says to me that we have to sign them.
0:56:49.680,0:56:54.720 ( some domestic situation itemised :) ) And then this is for getting the schema, that we loaded via the
0:56:54.720,0:56:59.200 OOBIs, that I showed you at the beginning, this is diagnostics we're taking a look at, the escrow
0:56:59.200,0:57:03.120 status, because all events are asynchronous, passing events around, you can get them out
0:57:03.120,0:57:06.560 of order, you can get them without full signatures, without full witness receipts, and we have a whole
0:57:06.560,0:57:12.560 series of escrows, that the events just sit in, that we constantly check for any other event to come in,
0:57:12.560,0:57:19.920 to resolve them out of their current state, this is for mailbox notifications, so this is a 'server-
0:57:19.920,0:57:26.400 sent events', streaming service for the agent U/I, to get notifications from the KERI system itself,
0:57:28.320,0:57:33.680 these are all the APIs for performing multisig and these are APIs for doing credentialing with
0:57:33.680,0:57:37.040 a multi-sig identifier. And the reason why these have to be different, is because there's
0:57:37.040,0:57:43.120 communication involved. Someone leads the process that would be with a POST, and then others call PUT
0:57:43.120,0:57:48.720 to participate in the process. And so, for example if I have an identifier, if you and I are sharing
0:57:48.720,0:57:54.480 an identifier, where you have a private key and I have a private key, I would initiate it and then
0:57:54.480,0:57:58.400 send, the system would send the credential to you, you would look at the credentials, say "Yep, this is
0:57:58.400,0:58:03.120 the one we agreed to issue". You would then do a PUT with that credential and your signature, and then
0:58:03.120,0:58:08.240 it would become a valid credential. And we do have scripts, and all that, we can go through that
0:58:08.240,0:58:13.120 Maybe you and I can schedule something next week, and we can go forward and do the multi-sig stuff
0:58:13.920,0:58:21.200 Okay, all right, so that's most of what I did. We had an hour and a half, when I was at
0:58:21.200,0:58:27.920 at IIW. Because I did it over the working lunch session (which meant I didn't get to eat) and we
0:58:27.920,0:58:32.800 went on through the next set of scripts, which is multi-sig and then multi-sig delegation on both
0:58:32.800,0:58:36.560 the delegator and the delegatee. And that was fun because you get four windows running, and they're
0:58:36.560,0:58:40.880 all waiting for each other. And when you're to finally fire off the final one, they all just resolve,
0:58:40.880,0:58:45.440 with signed events everywhere, It's really cool so maybe we can do that sometime next week. Steven: okay