- From: CCG Minutes Bot <minutes@w3c-ccg.org>
- Date: Wed, 11 Dec 2024 16:51:29 +0000
- To: public-credentials@w3.org
Thanks to Our Robot Overlords for scribing this week! The transcript for the call is now available here: https://w3c-ccg.github.io/meetings/2024-12-11/ Full text of the discussion follows for W3C archival purposes. Audio of the meeting is available at the following location: https://w3c-ccg.github.io/meetings/2024-12-11/audio.ogg A video recording is also available at: https://meet.w3c-ccg.org/archives/w3c-ccg-weekly-2024-12-11.mp4 ---------------------------------------------------------------- W3C CCG Weekly Teleconference Transcript for 2024-12-11 Agenda: https://www.w3.org/Search/Mail/Public/advanced_search?hdr-1-name=subject&hdr-1-query=%5BAGENDA&period_month=Dec&period_year=2024&index-grp=Public__FULL&index-type=t&type-index=public-credentials&resultsperpage=20&sortby=date Organizer: Harrison Tang, Kimberly Linson, Will Abramson Scribe: Our Robot Overlords Present: Christopher Allen, Harrison Tang, Alan Karp, Mahmoud Alkhraishi, Vanessa, Greg Bernstein, Tim Bouma, Manu Sporny, cc, TallTed // Ted Thibodeau (he/him) (OpenLinkSw.com), Will Abramson, Nis Jespersen , Tom S, Jeff O / HumanOS, Joe Andrieu, Geun-Hyung, Erica Connell, Alberto Leon, Phil Long, Alberto Leon(BKC at Harvard), Benjamin Young Our Robot Overlords are scribing. Harrison_Tang: Good morning uh good morning everyone so welcome to this week's w3c ctg uh meeting so today uh we're very excited to have Krista Christopher Allen here to uh talk about Edge identifiers and clicks uh but before that just want to go through some administrative stuff and then also uh we are in the process of the election of a new uh ccg culture role so I'll talk about that a little bit later. Harrison_Tang: Now I just want to quickly start with uh a reminder on the code of ethics and professional conduct I just want to make sure that we hold the constructive and respectful conversations here. Harrison_Tang: A quick note on the intellectual property anyone can participate in these calls however all substance contributions to any ccg work guidelines must be the member of the ccg with 4 IP agreements signed so if you have any questions in regards to uh getting a w3c account or the w3c community contributor license agreement uh please feel free to reach out to any of the cultures. Harrison_Tang: Please note that these meetings are automatically recorded and transcribed and we will publish the meeting minutes and the audio and video recordings uh in the next day or 2. Harrison_Tang: We use our GT chat to cue the speakers during the call. Harrison_Tang: Cost to take minutes so you can type in Q Plus to add yourself to the queue or cue minus to remove. Harrison_Tang: All right I just want to take a moment uh to uh for the introductions and reintroduction so if you are new to the community or you haven't been active and want to re-engage feel free to just unmute and uh speak yourself. Harrison_Tang: All right next announcements and reminders any uh new announcements or reminders of the upcoming events. Manu Sporny: Um yeah uh just 1 uh heads up to the group um I'm trying to find. <harrison_tang> Add "Topic: <Edge Identifiers & Cliques>" Manu Sporny: Appropriate email um uh there is uh so we've done a number of presentations on uh BBS in this group so it's a new privacy preserving cryptography new as in 20 years old at this point um but there's a big push to get it standardized at uh at ITF um and uh uh. Manu Sporny: And so there's been a lot of discussion and work so um Greg Bernstein's done a lot of presentations on BBS and how it's useful for privacy preserving uh credentials uh work has been going on at the internet engineering task force uh specifically uh really in the crypto Forum research Group which is where new types of cryptography tends to be standardized um in their uh was repeated requests to. Manu Sporny: Adopt um uh these these specifications so BBS is already adopted it has cryptographic review but there are 2 new features that we're really looking for in a complete solution to enable and linkable uh credentials uh and so the official call is out now and it is only open until December 20th so it's a very short period where you have to write in and uh signal your support for the specification. Manu Sporny: Uh if you care about. Manu Sporny: https://mailarchive.ietf.org/arch/msg/cfrg/H3LfT7mrelFnDUvJ3kL6CNev3C8/ Manu Sporny: Privacy preserving cryptography uh and specifically kind of the BBS stuff that we've been you know working on for years here um this is an important uh Point um to go in and and uh write in your support I put a link into the chat channel uh if you go there and you respond to that email uh unfortunately you do need to sign up for the CFR Gmail mailing list for your email not to be bounced um if you responded that and say that you supported adoption that's really all we're looking for now so again if you'd like to see this type of privacy preserving cryptography uh standardized and used and verifiable credentials and dids and all that kind of stuff uh it is vital that you go and um provide your support uh to the call for adoption that's it. Will Abramson: Yeah I just wanted to ask my that's different to like the past like I feel like I've already done that but like it seems like this is a new thing. Will Abramson: Ah okay so this is more important yeah okay. Manu Sporny: I know it's confusing the the previous 1 was a request that they start a call for adoption this 1 is the actual call for adoption which is. Manu Sporny: Kind of ridiculous. Manu Sporny: You know whatever so please write in again reaffirming that you would like the things adopted that's it. Harrison_Tang: any other. <greg_bernstein> Thanks all from the BBS team! Harrison_Tang: Announcements or reminders. Harrison_Tang: Just a quick previous what's coming so uh next Tuesday we'll have uh Isaac to uh give an update on the verifiable issuers and verifiers and then we will be uh not have we will not have to uh ccg meetings on December 24th and the December 31st and we will resume on January 7th uh with. Harrison_Tang: Presentation by JC. Harrison_Tang: I will send out the email reminders as always. Harrison_Tang: All right any uh last calls for the announcements and reminders. Harrison_Tang: All right so I just want to take a moment uh to open up the to talk about the election of the w3c ccg culture position so uh. Harrison_Tang: So the nomination has closed and we have 2 candidates uh Jillian Walsh and Mammoth uh I'm not gonna butcher his last name but we'll have 2 candidates uh and then I would like to invite them to actually share a few words and then uh right afterwards I'm going to send out uh an email uh to open up the voting and then we will have we will have the about a week uh to for for the voting to happen and we'll. Harrison_Tang: So my mom do you want to uh just uh speak a few words about your candidacy. Mahmoud Alkhraishi: Um hi my name is Muhammad Al keshi um over the past few years I've had the pleasure of working with many if not most of you. Harrison_Tang: Thank you Mamou. Mahmoud Alkhraishi: And it has been a wonderful time uh I'm looking to be the nominee for co-chair and to do that I would love to be able to help push forward the ccg in a lot of the different ways 1 of them is I've been very happy with the diversity inclusion that's been going on I would like to increase that there are a lot of changes that have been happening the last 8 months to a year in the ccg especially on how the topics are picked especially on what kinds of things that are being talked about these have been wonderful and I would like to keep pushing that forward um I've worked with a lot of you on many things at here at the VC working group at the did working group at ITF or at Dak and I feel like I would be able to provide uh you know a different perspective and I think that would be helpful thank you. Harrison_Tang: And I don't think I see uh joining here but I think she has sent out an email about her around the issue so uh you can just refer to her email. Harrison_Tang: All right a quick note so I will send out an email for the. Harrison_Tang: Regards to how you can vote uh but uh I will just quickly show what that voting interface look like so give me a second. Harrison_Tang: I think people people get to see my screen right. Harrison_Tang: Cool all right so it is a rank Choice voting so the we have 2 candidates here so it takes like about 30 seconds to do it's really easy so just uh type in your name and email and then you'll get to this screen and then you just uh do the rank Choice and then that's it so. Harrison_Tang: Let me see okay so the nomination uh has closed uh the end of day yesterday so we will open up the voting uh I'll send out the email right after this and then the voting will close on the. Harrison_Tang: Uh the end of the day next Monday and we'll announce the election result uh during the ccg meeting next Tuesday. Harrison_Tang: And then the new culture term will start on January 7th the new year. Harrison_Tang: All right pretty straight forward um. Harrison_Tang: All right before we get to the main agenda um any last introductions reintroductions announcements or. Harrison_Tang: Work item updates relay stuff. Will Abramson: Yes I will say something so we've I've been working at it with my mood and a bunch of other people who are on the traceability um. Will Abramson: https://www.w3.org/community/reports/credentials/CG-FINAL-traceability-interop-20241204/ Will Abramson: I think it's like a task force right to publish a final report I did mention this last time but we finally managed to get them published so I guess you know we can celebrate that a little bit I just dropped 2 links into the. Will Abramson: https://www.w3.org/community/reports/credentials/CG-FINAL-traceability-vocab-20241204/ Will Abramson: Um documents like these are now like final w3c official Community reports. Will Abramson: Maybe I'll ask my mood if I have to say anything about the documents themselves but before that I just would invite anybody if you're working on a report or if you have a report that you think is a final Community report but hasn't got this like official w3.org URL link yet like please reach out to me and and I'll work with you to help get them to this sort of more official stage. Will Abramson: So I think. Will Abramson: Get a proper you. Will Abramson: And we can put it on our community uh. Will Abramson: Stuff you know just mix things a bit more official. Will Abramson: So I don't know I moved you on to speak a little bit to these documents like. Mahmoud Alkhraishi: Yeah happy to walk through them um I want to say thank you all this I know this is not an easy process there's I know we haven't done it in a while and there's a few hiccups um. Mahmoud Alkhraishi: I wanted to have a chat with you afterwards maybe about a few things we could do to hopefully streamline this for the other participants but um there are 2 reports 1 for the traceability vocab 1 for the traceability interop both have been a work in progress since uh 2020 and they both tackle to close topics but they're a little bit different the traceability vocabulary aims to provide a vocabulary that would be used for asserting claims and verifiable credentials that talk about Supply chains and traceability within them they talk about products they talk about organizations they talk about chemical properties they talk about attributes uh like uh country of origin and that kind of stuff things about the Providence of a uh credential the traceability anthrop is started off as a profile of the VC API that was a little bit more restrictive with a few different um. Mahmoud Alkhraishi: Things picked out so that it would Aid in Enterprise uh interoperability that was specific to the traceability sector so there were a lot of assumptions about the kind of company that would run it there's a lot of assumptions about the kinds of things that you would want to do it and so we aimed to keep it in line with the VC API for the first while um after I want to say 3 years it started becoming its own thing and uh now it's published and it is intended to be used as a way to provide traceability documents from 1 point to the other over long-lived workflows that would allow people to correlate that would allow organizations to correlate credentials that relate to a life cycle of a specific product so there's a lot of things to go over there I'm happy to go in depth at any time anybody would like but um at risk of hijacking this meeting if there are any questions I'm happy. Harrison_Tang: Thank you thank you will and thank you Mammoth and by the way I probably would like you to uh uh come here and present those uh 2 papers uh at some point uh in the next the in the new year. Mahmoud Alkhraishi: Answer them yeah. Harrison_Tang: Sounds good thank you. Harrison_Tang: Last call or the announcements reminders or what kind of related thanks. <christopher_allen> First post is: https://www.blockchaincommons.com/musings/musings-cliques-1/ Christopher Allen: https://www.blockchaincommons.com/musings/musings-cliques-2/ Harrison_Tang: All right so let's get to the main agenda so again we're very excited to have Christopher Allen here to talk about Edge identifiers and clicks I saw that uh I saw his blog post about a month ago and I thought it was really really good uh especially I did have the same questions in regards to the single key Paradigm and uh I think he in his block goes he actually uh correctly pointed out in my opinion uh the flaws and the challenges in regards to the single click uh Paradigm how we can actually uh change that using Edge identifier so without further Ado um Christopher Allen uh your the floor is yours. <mahmoud_alkhraishi> as a correction, i believe the traceability wokr started early 2021 not 2020! Manu Sporny: You're muted Christopher I think. Harrison_Tang: Yes sorry of course the firm. Harrison_Tang: I think you're on mute still. Christopher Allen: Can you hear me now. Christopher Allen: Okay so um uh. Christopher Allen: I um uh have been talking about this for a while I think the first time uh I did it was at the last TPAC a little more a little uh uh briefly um I also have a brief video that I'll share um uh at the end of the meeting which is a a longer version of what I'm going to go through here um but what I'm really wanting to do with this community is talk about the implications of this so let me first reiterate it so that people have the basic understanding and then um I can go into some more uh uh specifics so um uh you know if you don't know me I've been involved for a very long time in this community and then before that uh you know with organizations like digicash and pgp and whatever uh uh uh prettl and then I was the co-editor and co-author of the TLs protocol and then blockchain Commons is. Christopher Allen: An organization. Christopher Allen: On uh interoperable secure and compassionate infrastructure uh in particular decentralized Solutions where everybody wins um. Christopher Allen: So as. Christopher Allen: Uh Harrison mentioned there is the uh uh the single signature Paradigm if we look at you know uh digital signatures today they were uh you know invented in uh kind of a a rough form in the late 70s really in the uh 80s is when the cryptography uh emerged and also a lot of the patents started getting locked up um and uh uh you know you know with real deployment of them in the 90s uh in Technologies like pgp early on and then later uh SSL TLS uh uh. Christopher Allen: Etc to uh you know move to today and to a certain extent we've been locked into a lot of the assumptions of that basic technology um uh and I think that there are now real solid mature and uh uh you know opportunities to rethink that and this is just the beginning of a discussion about that so um uh let's first talk about the the the the dangers of the Paradigm obviously you have a single point of compromise you know whether or not it's uh you know somebody stealing uh a key or uh doing a uh uh supply chain attack or um you know dealing with side channels Etc it's just a real uh uh. Christopher Allen: You know. Christopher Allen: Uh risky single point it's also a single point of failure in a variety of ways that aren't about compromise you know some of this is you know because of key fragility because of bit rot you know I have some old pgp keys that I can't seem to get to work anymore sometimes it's format changes you know I have some some uh you know uh you know old uh keys from the early days of Bitcoin that I can use but I have to keep an old copy of the software around. Christopher Allen: And then you know as we started doing a lot of the work in uh dids uh we really said hey key rotation and key separation are really important and it's hard uh to do with the single signature par Paradigm and then I wrote a book with Shannon and 1999 that talked about the 28 adversaries of keys and this is the the list categorized in uh you know these uh uh 7 categories um and this was really focused on digital asset Keys uh since that book I've you know at least added uh 2 or 3 uh possibilities for um identity focused adversaries uh for Keys um so we've got to address these problems uh and um uh now they're begin to be some new opportunities to do so. Christopher Allen: You know where I really want to start with and I I can't say I'm the first 1 is is uh you know talked about this but I think cryptographically it's been hard to do um and now we have an opportunity to do it so um if we go back to uh you know my original vision of self Sovereign identity it really was all about your relationships uh you control your identity but you don't control the network and we you know it was designed the whole point of the 10 principles was to support human dignity but also to allow people you know people to be peers not petitioners um and uh you know so but still it's not like you're this isolated uh being in this Digital Universe you have relationships and um uh we can't forget that in in SSI um so. Christopher Allen: So uh if we talk about relationships from a sociological perspective you know identity is actually decentralized already it's a bunch of edges you know when um you know your first identity was your you know mother's son or daughter. Christopher Allen: Uh you know while you know you were still in in her um uh pregnant body and um. Christopher Allen: That it was that is in effect the first you know identifier uh and that is a uh you know affiliate link and uh you know which goes backward in time forward in time but we also going to have fraternal links Etc so we can kind of consider these edges as membranes where there's kind of a a selective information you know uh between the entities you know um and there's a lot of you know papers if you look up local names or pet names Etc uh that have talked about you know how this um uh can be a different way of approaching identity um. Christopher Allen: Now we come to you know some modern technology Shore is a particular signature um uh uh system that I admire and 1 of the things is that some of the modern advances leveraging Shore have the power to create these relational relational edges so for instance 2 parties can create a key pair together where each party contributes a a secret but in fact the private key only exists in kind of what I call a cryptographic fog um and 1 of the interesting things about Shore is that these multi are the same size as a single Sig they are you know you know unless you reveal that they are a multi-sig you know they're indistinguishable from a single signature um and this means in the that the the group public key of these 2 parties uh is um. Christopher Allen: Is an edge. Christopher Allen: I mean it's a public key it looks like a public key of an individual uh but in this particular case it is referring to say the edge between Anna and Mary and Mary and Joshua um this you know group sort of fog private key allows for joint signatures so uh you can sign things together uh either indistinguishable from you know just being the the the that edge or you can say you know Mary and Joshua have done so together but it's a single uh cryptographic uh proof um so uh that really kind of leads up from edges to uh groups so um let's talk about a closed click which has some you know interesting opportunities uh this is the simplest form of a click um where there's basically an edge identifier between every pair of entities um and then together these married. Christopher Allen: To Joshua. Christopher Allen: Joshua um uh identifiers and keys can together create yet another key which identifies the click and that means that all 3 parties can participate in joint decisions and signatures and once again be indistinguishable and not correlate with the individual parties um uh and then of course you can go up forward from there you can have higher order clicks uh you know try to assist the simplest but you know it's um uh you know a uh an a. Christopher Allen: I'm not. Christopher Allen: Ally which it is it's not it's not a uh it it isn't a straight graphic its larger the more members there are the harder it is to close the graph um but you know you can do some fairly complex things here um and you can do these really interesting uh. Christopher Allen: Things where Clicks in turn can be recursive so you can have a click and another click and another click which together create an even larger group uh a click of clicks um and I think this also offers some very interesting power that Central signature and that diagram again is indistinguishable from the signature of uh uh a single person um. Christopher Allen: So what are the advantages of these Edge identifiers and clicks well um among other things in decentralized identity management it it offers some opportunities for peer-based identity creation and uh also peer-based identity validation in ways that um uh don't require uh uh a centralized party it's just the peers that do it um the uh there is uh uh Spock and Spa are are single point of compromise single point of failure um resilience so distributed control guards against different kinds of compromise and fail failure we have opportunities for secret group decision making and we have the opportunities for enhanced privacy and there are a lot of different variants here I'm not proposing any particular 1 here I'm just saying there are new opportunities uh given this um there are uh some identifiers uh click identifier drawbacks they are. Christopher Allen: Are more. Christopher Allen: Complex right now they require um you know multi-signature technology which has you know matured in the last 5 6 years uh but is new um multi-set instantaneous so there's no instant gratification and it is a new paradigm you know there are you know going to be new risks that are uh you know as we create these systems that you know we've not seen before um. Christopher Allen: Very quickly go through this page uh uh so obviously you know those clothes clicks are hard they get bigger and bigger uh but you don't necessarily have to have closed clicks you could have open clicks which is often more realistic um uh they lose some graph analysis advantages and certain cryptographic things you can do when a click is closed but it also you know offers some new possibilities. Christopher Allen: I particularly like what I've been calling fuzzy clicks again you know my own terminology for it but um uh they allow for threshold signings so that just some members of The Click uh can participate so uh you can be in say in a 2 of 3 it can be um you know Mary or Bob or Bob or Joshua or Joshua and Mary uh that can uh sign for the group um. Christopher Allen: And uh but I think the real opportunity here is that clicks can be used with devices so devices can be part of clicks uh you can have your ring and your phone and uh you know an offline Service uh together be part of a clique that is a 2 of 3 and if any 1 of them fail uh you still have your identifiers and your keys and that what you need to do to rotate and move things forward um and then of course these can be combined with the you know person oriented clicks so um now with this uh Triad uh that is the relationship well it's really a a diad of Mary and Joshua uh but if Mary lost her secret or Joshua lost his uh they've lost their uh their Edge identifier but if they do it in a click form where the third party is a uh. Christopher Allen: Uh you know a hardware click they have now uh increased reliability um either 1 of them can fail and then the the remainders can basically recover uh Keys uh rotate Keys Etc. Christopher Allen: So let's talk about the uh the implications for identity so um you know if we look at uh you know this is really focused on you know the kind of work that we do uh if we look at the the classic faux clear uh peer claim uh and you know I know there are lots of uh limitations of faux and whatever but this is an example of 1 where you know Christopher a says he knows the public key of wolf um and is verifying it uh in the in the original faux uh you could the uh there was this concept of uh uh of um uh nose being this 1-way relationship but then they also had this concept of up here well you couldn't really do the peer uh directly you could maybe have wolf know the public key of Zid you look at these 2 I mean the did ish type uh uh object um uh you could they could basically say in return I know. Christopher Allen: The wolf knows. Christopher Allen: And then you can infer that their peers this isn't really a peer statement as opposed to uh in a uh uh an edge credential and kind of a VC style you know we are a peer group uh here are the 2 peers here's the public key that is the public key of the of the peers you can get the public key of Christopher a the public key of wolf aggregate them and go oh yeah that is the public key of the of them both and then verify it so that gives us a lot of you know interesting opportunities to make crypto ref proofs about these types of things um and of course with Elysian uh you know we don't have to reveal that those are actually 2 parties so you we can basically say this is a peer group but we're not saying who the peers are but you can still validate the signature um uh so I think there's some real you know interesting opportunities uh in that uh in that. Christopher Allen: In that space and I've been exploring this a lot more um with different kinds of proofs and you know uh when might you want to align signatures or signature um metadata uh also you know trying to puzzle out how to do this with object capabilities and uh some other opportunities. <phil> Does the ELIDED (2) convey that the group consists of a pair? Christopher Allen: So some final notes uh the single signature Paradigm is not enough we really need these relational identifiers for peers and groups um there are the interesting exotic opportunities in nested open and fuzzy uh groups they're interesting challenges um and this is a really rich uh Paradigm uh the uh the musings uh uh number 1 is the uh the first article on this topic um I will put that link in the chat in just a second. Christopher Allen: Me uh stop the. Christopher Allen: Okay so uh let me share my window other window. Christopher Allen: There is the share button now. Christopher Allen: There we go. Harrison_Tang: It's the third 1 at the bottom. Christopher Allen: Okay so um blockchain Commons is my main website uh the first Edge identifiers uh. Christopher Allen: https://www.blockchaincommons.com/musings/musings-cliques-1/ Christopher Allen: Uh post is here I'll put that in the chat. Christopher Allen: Uh at the uh uh end of that post is uh a link to uh the open and fuzzy uh follow-up uh there's another version of my presentation on YouTube um the you know we have a whole page on Music 2 technology and how it works including you know some of the implications of you know how you would do a sequence diagram to create this multi-sig between uh 3 parties um. Christopher Allen: Does not dive into snore before I have a whole bunch of stuff here on uh you know kind of making uh uh you know what is snore and how does it work um uh but I also have this thing called snore in a nutshell which is kind of an 8-bit uh version of snore so you can understand how how it functions uh we have a whole bunch of stuff on Frost which is uh the threshold algorithm we've held 4 know yeah 4 Workshop uh these are results from the first 3 Workshop we just held a workshop last week and this page will be updated tomorrow uh but there is a uh a frost meeting tomorrow and then we also have a lot of uh you know research papers in our uh blockchain Commons research uh exploring some of the the uh the challenges and stuff. Christopher Allen: Okay so uh I'm going to stop sharing my screen and. Christopher Allen: Uh open it up for questions. Harrison_Tang: Phil I think you have a question in regards to evision. Phil Long: Noted that when you showed the the way in which you could um share the the uh privately the the combined key and noted Illusions parenthesis 2 parentheses was that conveying that there were in fact 2 members of the group or was that meaning something else. Christopher Allen: Um so uh I don't want to go into detail on uh uh gordian envelope uh the. Christopher Allen: Uh but basically you have a choice of what you how you want how you want to structure these and what you wish to allow to be correlated so yes you can reveal that there are 2 entries there or you can choose to not uh and just say that there's a single entry there and and it you know because it's all a a recursive nested structure uh in this particular case I just simply because it is a reclaiming its appear I felt well you kind of already know it's at least 2 people. Christopher Allen: And uh that that was a better representative of it but that's actually been a lot of what I've been focusing my work in the last 2 months on is exploring uh you know what does it mean to Allied some of this different types of data especially you know with key separation where we might have a lot of different kinds of keys that have different uh uh capabilities different kinds of proofs including non-s signature proofs when would you want to Allied. Christopher Allen: Some portion. Christopher Allen: Or some details of them and I think there's a lot of work that that is required for that and I and you know have discovered a number of things in uh in uh kind of the the the Json LD style proofs that are kind of difficult to do um you know you know presuming we add a lesion to Jason just straight up Json LD uh and there have been some old proposals for how that might work uh it turns out it's actually uh some stuff is uh a little more hard to do um with some of the assumptions in in Json LD so I've been doing almost all of my stuff right now in uh gordian envelope which I you know maybe someday could be a an alternative to in say a D20 or a vc20 uh a core based a liable structure um but that's you know uh a ways away I'm more in the experimentation phase. Christopher Allen: Thank you. Phil Long: Thanks that's very helpful. Alan Karp: Uh yeah can you walk through how you would as an individual use these keys to to do something. Christopher Allen: Sure so um let me go to um. Christopher Allen: Let me go to this window. Christopher Allen: So this is Music 2 uh so Music 2 is uh an N of n uh protocol so you have to have you know it's not a threshold uh protocol um in its uh in its basic form and um uh you know from the you know uh the the point of view of Alice and Bob uh they're basically you know create kind of create some session keys for each other they're going to share some partial public Keys uh uh which will then allow them to uh to create a a group public key and that's what they share with the verifier um so uh then when they want to sign something they're basically doing this non exchange uh and then you know creating these parcel signatures which are basically combined which then allows you to have the the the final verified signature so when we look at this this very. Christopher Allen: Uh you know by default sees that there is a public key and assigned message and it verifies like any other schnoor signature um even though it was created by Alice and Bob and this basically represents you know them as uh and a you know as a a uh the simplest Association a diad um and uh what what's kind of interesting about this is that it's not you know obviously you know you could do a verifiable credential that was um you know 1 was signed by 1 and 1 and then there's you know a co-signature by that but that co-signature is what I call a comp computational uh smart contract in the sense that you know there's some kind of or statement uh you know excuse me and statement in code that is evaluate signature 1 and says that's true and then evaluates Signature 2 2 and says that's true and because there is this and. Christopher Allen: Statement in between. Christopher Allen: Now they're both true uh in a logical and but that computational um uh or statement is uh very easy to attack whereas the the cryptographic and of Allison Bob is a cryptographic function and it's not just simply swapping out the logical and operator in some hack for a logical or operator it the mass just simply will not work uh unless it is an and and I think there's some real power in thinking about the you know the Futures where you have these cryptographic and not comp uh uh you know scripts whether or not simple logical scripts and or not or more powerful scripting functions I've talked about that in in a couple of papers. Alan Karp: So yeah well not really because uh I said as an individual so this means that to to use these keys I need to coordinate with at least 1 other party is that true. Christopher Allen: Yes um but that being said if we go back to um where is my. Christopher Allen: Okay so if we go back to this um you know this is 1 of the first things that we've been implementing here at blockchain Commons is uh you know 1 of our patrons is a is a ring um uh cryptographic ring and uh it has cryptographic material on it uh. Christopher Allen: A you know a phone that acts as a coordinator that has another um uh excuse me the Hub would be the phone um which has another uh secret on it um and all I need is my you know my ring and my phone which are together here with me to represent me uh and even though there is uh you know an operation that is happening here um the advantage of having the this in a in a 2 of 3 and a in a fuzzy click is that the um you know I can have a an offline or another uh you know social key recovery service or whatever that if you know I my phone dies or Android uh uh you know uh becomes corrupted uh that single key will not be a single point of compromise my my offline service and my ring can still recover you know my public identifier. Christopher Allen: Um so but yeah any 1 of these operations is going to require some kind of communication between uh 2 devices or 2 people or 2 services at minimum. Alan Karp: Okay know that answers the question thank you. Manu Sporny: Yeah this is a great Christopher I I read um either your blog posts and kind of tried to internalize how we could maybe apply some of this stuff to the existing did and in VC ecosystem um so I I think I get the the whole like you know to issue something you've got to run this cryptographic protocol to get the parties together to share you know parts of their key and generate the signature um for the verifier it seems like the only thing the verifier needs is is the combined you know public key um and the signature and they should be able to verify that just in and of itself is that correct. Christopher Allen: That is correct um so you know let let's talk about you know 1 of the interesting things I think we first talked about this in um uh when you presented verifiable credentials at our wat 2 in in New York City uh after the UN event which is there uh you know we have all of this work this desire to be decentralized uh but then we have this problem of the issuers are centralizing because in issuer is making a claim and that is a natural centrality and um uh however I think there's some interesting opportunities maybe not precisely decentralized but in more distributed Fashions uh you could basically say the DMV which among other things requires uh uh uh you know a uh uh driver's test uh you could have uh you know. Christopher Allen: Drivers test which is judged by multiple parties who uh and it's some Quorum of them which might be 2 of 99 uh to basically say yes this person has passed their driver's test and uh now you have this uh this proof uh that is you know represents the the the all of the DMV and there's no single point of failure um I could see also a lot of interesting Hardware uh things we can do here where um especially as we're exploring uh key rotation and key recovery right now there are very explicitly proven and working um key recovery things so if uh you have a bit rot on 1 of your uh Quorum devices uh that that key can be uh restored if you now want to move from say A you know 3 or 5 to a 4 of 9 you can create a new Quorum uh and still keep the same public key those are. Christopher Allen: I I considered. Christopher Allen: Be relatively mature. Christopher Allen: There are papers and other you know work beginning on also how to do approvable rotation where the the the old Quorum whatever remains of it you know uh say it's a 3 of 5 and 2 of them have gone down and they're going oh no we're now we're you know any 1 of us fail the whole Quorum fails well they can basically not just regenerate the 2 Bad Keys maybe they were compromised instead of uh just bad uh they can basically create a new public key identifier plus a cryptographic proof that this could not have happened if the first Quorum didn't exist um and then once again we're back to a uh a new um uh you know a a real cryptographic key rotation uh and thus eliminating and distributing some of the risks of the single keys in uh in an issue work. Harrison_Tang: Well no I think you have a follow-up question. Manu Sporny: I do but I'll I can go after well. Will Abramson: Cool yeah thanks I wanted to say 2 things the first is kind of uh. Will Abramson: Similar to on my building on my manufactured like in theory you know if there was a snow sep uh verification like signature Suite that existed today which you know maybe doesn't I would like that to be people could be issuing credentials using music or Frost and from a verification perspective you wouldn't know like you know like it would just be the same process the signature would look the same you'd be able to verify a credential exactly the same this this whole stuff that Chris was talking can I happen before we create the signature it's like some internal you know it's kind of like. Will Abramson: Did want to. Will Abramson: Better support it but like we need to other specs like the bids are kind of already could support it if if people could create these signatures. Will Abramson: Uh and I think you spoke to I just went to like this very clear right we could do this today we just. Will Abramson: Don't do it. Christopher Allen: Correct um you I don't believe that there are any current um uh. Christopher Allen: Uh Signature suites oh wait wait wait there is 1 hold on let me let me show you uh let me uh share window. Will Abramson: Well I thought Ed 25549 was maybe snore compliant or like similar to. Christopher Allen: Okay so it is and it isn't so let's be clear here um so this is 1 of the more mature uh implementations at this point it was found uh I presume are you seeing my screen. Christopher Allen: Okay so this is the zcash foundation they funded the frost 1 of the major Frost libraries and they also funded the uh uh the development of an ITF uh uh research RFC uh that uh is just on the signature part of the frost uh functionality um and they do have a frost ed25519 so yes you can uh do frost uh uh with 255125519 and generate a key uh there are some interesting implications of that this is not really completely compatible with. Christopher Allen: Other 25519 um uh implementations because there are bugs in 25519 um there are a lot of conformance problems between different libraries of 25519 so if you tried to use uh I mean in fact these days and we just did a 25519 implementation for um. Christopher Allen: 40 And envelope and we basically had to find a library that reproduced the bugs of the SSH. Christopher Allen: Cryptography um there are also other problems with uh ed25519 uh in the fact that uh uh that some of the aggregation capabilities of it um are don't work because of some choices that were made uh early on to avoid side Channel attacks um so the most of the the uh to avoid those there is something that is very close to Ed 2551 25519 called ristretto um and it doesn't have either of these problems um but again you know that's not going to be conformant uh you know to um to uh you know SSH uh implementations of 255519 um there's a wonderful I I I I'm not going to try to pull up the the the the the uh the document here um but there is a. Christopher Allen: Ah uh. Christopher Allen: Uh you know a paper that sort of describes you know the the the challenges there and there are a number of cryptographers that are kind of going hey you know we really need to respect to Phi you know completely redo 25519 because it you know because of these uh. Christopher Allen: Valuability and other problems um. Will Abramson: I do think I suppose we had like a snore sep uh signature Suite like it would be interesting to think about like what would it actually have to add to that to like support to support like creating these types of signatures I think it would just be an extra sort of paragraph in the create proof section that says you could also create your approved using this algorithm just point out to the you know like here's it described over here like the the like ccg or the work items here wouldn't be defining that stuff just it's defined over here as long as you can create a proof following this process stick it in the DC. Will Abramson: They should be verifiable using regular small verification. Christopher Allen: That's correct and I mean basically if you I mean if as far as the VC is concerned the spec and whatever it's just a single signature like any other signature that but in fact behind the scenes is uh you know a more complex uh proof so yeah you can start using this stuff now this is kind of why I've been puzzling around this whole issue of this metadata around signatures um you know how how do you include the fact that there are other things going on behind the scenes you don't really want to put it in the the vert the in the the triples of the verifiable credential it really kind of belongs in the signature block to basically say hey there's other things that were happening here if you need to know them here's how you might. Christopher Allen: You know. Will Abramson: Yes if you wanted to tell someone this signature was created by 3 of 3 of 5 right and like have them be able to verify that themselves. Christopher Allen: Correct correct um so they can verify the signature trivially um it's you know how do you add that extra information and how do you secure that extra information as well as you know been 1 of the the challenges that I'm working on. Will Abramson: Uh I did have 1 more comment but Mana you can go no man can go. Harrison_Tang: I'm sorry. Manu Sporny: Thanks well yeah so adding it to the signature at least with the data Integrity signatures is pretty trivial like you can add arbitrary information to that and it is in the signature block and it is signed over a by default so that's a fairly Light Lift I would imagine that you know in in the worst case we're talking about creating another crypto Suite that is as will said effectively a copy and paste of the existing 1 with some minor modifications in there like for example the the Ed ed25519 crypto Suite that we have right now um does try to fix some of the malleability problems with ed25519 so so we we went in and and applied some of the you know the the best practices for making sure that ed25519 was a little more locked down I think the only other things we'd have to add is like will said like maybe a paragraph or 2 to say that you know if you want to create you know a frost based signature this. Manu Sporny: Is how you. Manu Sporny: Um we don't really even need to use ed25519 we could use ristretto or something else like that I think that there are 2 big challenges that I see here so I I don't see the crypto Suite as a challenge that's fairly uh uh simple to put together um. Manu Sporny: And it it ends up looking. Manu Sporny: Just like. Manu Sporny: Regular verifiable credential you know at the end of the day so that's that's a good thing um and that's not difficult um the the thing where I'm I don't quite understand how we we scale is you know for for every click. Manu Sporny: You have to kind of identify that click I mean that's and the clicks kind of identified through a public key but but then how does the verifier know whether or not to trust that click my my presumption here is that you'd put it in something like the did document you would just list it as another key that can make assertions in your did document and the verifier somehow uh finds out that you know oh you are you know presenting this or or they they figure out a way to find you know that key publication in a did document that they trust and then that's the thing that gives them kind of trust in the issuing click right they they have to kind of do some kind of Discovery process to figure out who if if 1 of the issues issuers they trust is a part of a clique that ended up um. Manu Sporny: Issuing the the credential so do you have um what are your what are your thoughts on oh sorry that was 1 of the things so I I it feels like there's an explosion in public Keys when we need to put them somewhere like in a did document the other difficult thing as you know Christopher is like getting some variation of this through the ietf crypto form research group can take years um but this feels like. Manu Sporny: You know doable like I mean it if we started today I would imagine we'd be able to get this done in like 2 years um. Manu Sporny: So so what are your thoughts on kind of the standardization that would need to happen with some of this Frost stuff and then what are your thoughts on you know how does the verifier trust the issuing click. Christopher Allen: Right so I think it depends on things so I'll give you this example uh I I'm sure are you seeing my shared screen. Christopher Allen: Okay so obviously this requires nothing but you know just simply saying here's you know ah you know seventh uh um uh you know signature Suite which supports say uh you know a bip 340 Shore or ristretto 255 ristretto or whatever um uh you and this is again 1 you cannot do with 25519 um but this actually here is an interesting thing because when I reviewed that this is a peer group. Christopher Allen: I can simply do a a cryptographic finite field edition of this public key and this public key and it will equal this public key okay that's 1 of the things that schnoor offers is this finite uh field um uh aggregation capability so you know if you know who's what Christopher's public key is and what Wolf's public key is you uh then. Christopher Allen: This fairly trivial operation uh to determine yes that they you know this would not be possible unless those 2 were were the same and so you might verify the signature uh that is you know using this thing here but this sort of allows for some additional information revealed and I think this is a uh you know this is a small example of a whole bunch of these kinds of problems because for instance um uh this is what I call an accountable uh signature in the sense that um because we're using Music 2 uh we can through a variety of mechanisms know definitively that Christopher was 1 of the parties that signed it meaning Christopher is accountable as 1 of the parties in this in this Quorum of 2 of 2 or if you're using Taproot there's some other techniques where you're combining these um and if you go to my last frost presentation uh from last Tuesday um on YouTube. Christopher Allen: You can see. Christopher Allen: An example of it. Christopher Allen: You can know but Frost is not an accountable protocol in fact um you know even though I have secret material I can't prove that I contributed to the signature. Christopher Allen: Um I a quorum of parties can prove can prove that I contributed or make a claim and prove that I contributed to the signature but it's what is known as a non-accountable signature so it turns out there's some interesting advantages for having non-accountable signatures that you know prevent certain kinds of coercion allows for certain kinds of voting and anonymity but sometimes you want parties to be accountable so how do you do both is kind of 1 of my big challenges you know uh this month the next month um and you know I've begun to work on it in uh this uh research paper so and we've you know barely begun um. Manu Sporny: Yes very helpful thank you. Christopher Allen: Oh and 1 thing I was subtle in there because I hadn't really come up with it was in that 1 I'd actually put that peer information in the in with the credential um but in fact it really should have been in the signature block um and because but it's something extra in the signature block that requires it or the proof block uh it's some other proof that if you need that if you really want you know to verify that it truly is a peer because there's because that's important for some reason or a click or 1 of these other forms uh you know which can be a proof of of Click formation you know you can do all these social you can do all these graph cryptographic graph things with it um you know so you the signature may not be all you need you also want in some cases to do this other stuff it really belongs in the signature block. Harrison_Tang: Thank you thank you Christopher I always enjoy when you uh jump on and talk about this kind of stuff so thanks a lot. <christopher_allen> My email ChristopherA@LifeWithAlacrity.com Harrison_Tang: All right so who's uh this week's uh ccg meeting uh I have to send out the email in regards to the culture vote uh if you have any questions just uh email me back but thank you thanks Christopher thanks Manu thanks will thanks everybody uh today. Christopher Allen: Okay and my emails in the transcript.
Received on Wednesday, 11 December 2024 16:51:37 UTC