I just wanted to voice my opinion that I like today's program and the topic a lot. I would be there but unfortunately I have two other working groups that I need be at. But I'll be following your discussions after the fact, and also want to contribute in some way. Have fun!
", "time": "2022-11-07T15:43:09Z"}, {"author": "Tobia Castaldi", "text": "test
", "time": "2022-11-07T15:45:02Z"}, {"author": "Dirk Kutscher", "text": "works
", "time": "2022-11-07T15:45:57Z"}, {"author": "George Michaelson", "text": "I'm in hedgehog taking notes, I have no idea if anyone else can see this
", "time": "2022-11-07T15:46:34Z"}, {"author": "George Michaelson", "text": "zulip refused to let me in, I cannot see any link said to be \"in the chat\"
", "time": "2022-11-07T15:46:48Z"}, {"author": "Dirk Kutscher", "text": "So far, https://notes.ietf.org/notes-ietf-115-dinrg is mostly empty \u2013\u00a0where are you taking notes?
", "time": "2022-11-07T15:47:16Z"}, {"author": "George Michaelson", "text": "in meetecho in the hedgehog app
", "time": "2022-11-07T15:47:39Z"}, {"author": "George Michaelson", "text": "I can cut and paste into the one you want
", "time": "2022-11-07T15:47:46Z"}, {"author": "Dirk Kutscher", "text": "No, that's the same document.
", "time": "2022-11-07T15:48:22Z"}, {"author": "George Michaelson", "text": "Is that visible?
", "time": "2022-11-07T15:48:27Z"}, {"author": "Wolfgang Beck", "text": "hm, layer1 is very much still decentralized, ownership is not
", "time": "2022-11-07T15:48:40Z"}, {"author": "George Michaelson", "text": "well.. can't explain why it shows as empty when I've been typing in it
", "time": "2022-11-07T15:48:44Z"}, {"author": "Dirk Kutscher", "text": "Seeing text now! :-)
", "time": "2022-11-07T15:48:54Z"}, {"author": "George Michaelson", "text": "@wolfgang depends what economy you are in
", "time": "2022-11-07T15:49:01Z"}, {"author": "George Michaelson", "text": "@wolfgang some economies have regulatory control for L1/2 and then competition
", "time": "2022-11-07T15:49:20Z"}, {"author": "Wolfgang Beck", "text": "@george what i mean is that the big ones like google still use decentralized links and network elements internally. So technologically decentral, but economically a huge centralized hub
", "time": "2022-11-07T15:53:52Z"}, {"author": "George Michaelson", "text": "@wolfgang true
", "time": "2022-11-07T15:57:04Z"}, {"author": "Andrew Campling", "text": "Agree on gatekeeper features in protocols, especially if there is no provision for user control or selection inherent in the protocol design
", "time": "2022-11-07T15:59:56Z"}, {"author": "Wolfgang Beck", "text": "If IRTF solves the problem of economical centralization, we have solved world hunger.
", "time": "2022-11-07T16:01:49Z"}, {"author": "George Michaelson", "text": "this is a problem statement taxonomy not a solution at this stage
", "time": "2022-11-07T16:02:16Z"}, {"author": "Carolina Caeiro", "text": "it would be nice to see the draft reflecting on the impact of centralization on connecting the unconnected
", "time": "2022-11-07T16:02:28Z"}, {"author": "George Michaelson", "text": "@Caroline GAIA work?
", "time": "2022-11-07T16:03:23Z"}, {"author": "A.J. Beal", "text": "I'm confused about the distinction between \"Architecture Centralization\" and \"Server/Application Centralization\"? If a decentralized application on a decentralized topology makes transport layer optimizations -- I'd see it causing interoperability problems potentially, but it would still be a decentralized system no?
", "time": "2022-11-07T16:03:53Z"}, {"author": "Wolfgang Beck", "text": "I'm worried that the economical - that is social - problems are hard to solve with technological solutions
", "time": "2022-11-07T16:03:56Z"}, {"author": "George Michaelson", "text": "CDN == death of transit
", "time": "2022-11-07T16:04:21Z"}, {"author": "George Michaelson", "text": "but also makes addressing ephemeral, even transport protocol.
", "time": "2022-11-07T16:04:38Z"}, {"author": "Carolina Caeiro", "text": "@George -- it could considering articulation between the two
", "time": "2022-11-07T16:04:39Z"}, {"author": "George Michaelson", "text": "@carolina good idea!
", "time": "2022-11-07T16:04:52Z"}, {"author": "Wolfgang Beck", "text": "@AJ centralization should be discussed per protocol layer.
", "time": "2022-11-07T16:05:24Z"}, {"author": "Andrew Campling", "text": "Reflecting on the CDN question posed by our chair, of course some protocols are being developed now for use by CDNs - for example ECH
", "time": "2022-11-07T16:05:25Z"}, {"author": "David Oran", "text": "\"E2E encryption forces data to the endpoints\" - is that only true for channel encryption? Doesn't object encryption allow data to be anywhere?
", "time": "2022-11-07T16:08:40Z"}, {"author": "David Oran", "text": "Thought to ponder: Our focus has been focussed internally to the IETF, but what about focusing at least in part externally to inform anti-trust regulation/legislation/enfocement?
", "time": "2022-11-07T16:13:04Z"}, {"author": "Dirk Kutscher", "text": "Yes, that's one aspect of our re-chartering discussion.
", "time": "2022-11-07T16:13:37Z"}, {"author": "George Michaelson", "text": "If people don't say their names then the record is going to say ???? its like 25 years of IETF praxis is SAY YOUR NAME </grump>
", "time": "2022-11-07T16:14:38Z"}, {"author": "Christian Huitema", "text": "In the media-over-quic discussions, people are using terms like \"glass to glass\" rather than end-to-end
", "time": "2022-11-07T16:15:11Z"}, {"author": "George Michaelson", "text": "session layer hides interface bindings \"host\" was never the host in this address world
", "time": "2022-11-07T16:16:15Z"}, {"author": "David Oran", "text": "@christian: until we have neural implants in the optic nerve that can do crypto :-)
", "time": "2022-11-07T16:17:37Z"}, {"author": "Wolfgang Beck", "text": "I'd argue that Google's and Facebooks web servers are pretty much decentralized around the world
", "time": "2022-11-07T16:20:19Z"}, {"author": "George Michaelson", "text": "so is the thesis of this presentation \"REST as implemented is bad\" ?
", "time": "2022-11-07T16:20:55Z"}, {"author": "Marc Petit-Huguenin", "text": "POST creates, PUT updates.
", "time": "2022-11-07T16:21:24Z"}, {"author": "David Oran", "text": "@wolfgang: yes, but operated by a siingle administration, and in most cases carrying data over only private/internal communcation facilities
", "time": "2022-11-07T16:21:49Z"}, {"author": "George Michaelson", "text": "@wolfgang I would agree. eventual consistency in a distributed model is remarkably resilient to global failure at a cost of local failure and de-coherence into time
", "time": "2022-11-07T16:21:53Z"}, {"author": "David Oran", "text": "@George: as log as that's the tradeoff you wnt with the CAP throrem. Good for somethings but not all.
", "time": "2022-11-07T16:22:30Z"}, {"author": "Wolfgang Beck", "text": "@david yes, it's economically centralized. Is this an economical or a technical problem?
", "time": "2022-11-07T16:23:21Z"}, {"author": "Jim Reid", "text": "It would be good to document those trade-offs => informed decision-making.
", "time": "2022-11-07T16:23:41Z"}, {"author": "David Oran", "text": "@wolfgang: both, plus political
", "time": "2022-11-07T16:24:02Z"}, {"author": "Jim Reid", "text": "\"Is this an economical or a technical problem?\". Yes. It's both of these things (and others too: privacy, public policy, regulatory,etc)
", "time": "2022-11-07T16:25:10Z"}, {"author": "David Oran", "text": "For example, you could have gotten physical redundancy and viable decentralization for redundancy and resilience with simply co-lo facilities and multiple competing transit networks, but those lost out because of the superior economics and poor regulation of surveillance capitalism.
", "time": "2022-11-07T16:26:50Z"}, {"author": "Andrew Campling", "text": "It would be interesting to test whether any of the trade-offs / centralisation effects are in the long-term interest of end users, and whether any efforts have been undertaken to ask them via multistakeholder engagement (ie RFC 8890)
", "time": "2022-11-07T16:27:48Z"}, {"author": "A.J. Beal", "text": "I think a core issue in pushing decentralization is the current lack of support/reliability grantees for P2P applications. We've made serious strides with client/server models that have become ubiquitous - and that's driving centralization.
", "time": "2022-11-07T16:30:11Z"}, {"author": "Christian Huitema", "text": "How much is centralization driven by the financing of the internet with advertisement?
", "time": "2022-11-07T16:31:00Z"}, {"author": "Dirk Kutscher", "text": "A significant amount, I'd think.
", "time": "2022-11-07T16:31:38Z"}, {"author": "A.J. Beal", "text": "P2P rasies a lot of hairy questions we don't have answers to -- if we have some P2P web service where clients fetch web resources from other clients -- how can service providers push updated content? -- how do we efficiently solve reliable delivery at scale for P2P applications?
", "time": "2022-11-07T16:31:41Z"}, {"author": "Andrew Campling", "text": "@Christian it does seem to facilitate the surveillance capitalism model
", "time": "2022-11-07T16:31:44Z"}, {"author": "George Michaelson", "text": "@christian given the lack of evident competition, I would argue hugely
", "time": "2022-11-07T16:31:53Z"}, {"author": "George Michaelson", "text": "@christian not that competition guarantees no hegemony
", "time": "2022-11-07T16:32:06Z"}, {"author": "Christian Huitema", "text": "So what if personalised ads were banned?
", "time": "2022-11-07T16:32:29Z"}, {"author": "David Oran", "text": "@Christian: probably 80-90%, along with: (a)economies of scale that have not yet been challenged by serious dis-economies of scale, and (b)the unrealistic fulll mesh fantasies of the early e-commerce systems that got blown up by Amazon, EBay, and a few others.
", "time": "2022-11-07T16:33:21Z"}, {"author": "Andrew Campling", "text": "Or user tracking
", "time": "2022-11-07T16:33:24Z"}, {"author": "Jim Reid", "text": "Andrew [RFC8890 comment], that's as worthwhile as asking a goldfish about quantum physics. Most users have no idea how the interwebs work and only care that they get Netflix, Spotify, etc when they turn on the Internet. Come to think about it, I don't care which CDN feeds me.
", "time": "2022-11-07T16:33:25Z"}, {"author": "George Michaelson", "text": "@jim netflix cares about your password sharing and is monetising around it based on IP == identity
", "time": "2022-11-07T16:34:04Z"}, {"author": "Andrew Campling", "text": "@Jim we could start by engaging with civil society, industry and similar groups. Asking user does not have to mean asking consumers.
", "time": "2022-11-07T16:35:00Z"}, {"author": "Jim Reid", "text": "That's true George. The point I was trying to make was I don't (need to) know or care about which Netflix cache node throws bits at me.
", "time": "2022-11-07T16:36:44Z"}, {"author": "Wolfgang Beck", "text": "Skype had central control but P2P execution
", "time": "2022-11-07T16:40:01Z"}, {"author": "David Oran", "text": "@wolfgang: and they had more than a few dramatic correlated failures that took out zillions of users at once anyway...
", "time": "2022-11-07T16:41:13Z"}, {"author": "George Michaelson", "text": "this sounds like kerberos ticket-granting-ticket
", "time": "2022-11-07T16:42:51Z"}, {"author": "George Michaelson", "text": "I dont mean literally \"the same\" but the ontology of words here seem to seek similar functionality, we've been here before in some ways
", "time": "2022-11-07T16:43:15Z"}, {"author": "Wolfgang Beck", "text": "and moved to a more centralized architecture as a result. Wonder how it would have turned out if centrally controled hardware like Android/IOS phone had been a thing back then
", "time": "2022-11-07T16:43:15Z"}, {"author": "Britta Hale", "text": "@George It does indeed. However, if the user is doing a ticket granting service for themselves / their respective devices, then some of the typical Kerberos downsides do not apply. It is an intriguing idea...
", "time": "2022-11-07T16:44:16Z"}, {"author": "David Oran", "text": "PoA's are a nice abstraction al long as there is also an authenticated biinding to RWIs (real world identities) and the ability to sue for bad behavior by a delegate.
", "time": "2022-11-07T16:45:32Z"}, {"author": "Christian Huitema", "text": "Android is centralized software, not hardware. Software defines API, app stores, etc. Previous attempts (e.g. by Microsoft) have shown that it would be very hard to develop alternatives -- cannot get app developers to build for an alternative platform.
", "time": "2022-11-07T16:45:33Z"}, {"author": "David Oran", "text": "Ih addition to this \"signatory registry\" you also need the ability to audit actions by a different authority from the one you granted your PoA to.
", "time": "2022-11-07T16:47:09Z"}, {"author": "George Michaelson", "text": "@christian EU forcing the door open on software marketplace and signing of binaries on the platform. The idea of a gated s/w store is a poisoned chalice but so is breaking it down. we're going to get more bad binaries if this succeeds in opening.
", "time": "2022-11-07T16:47:20Z"}, {"author": "Wolfgang Beck", "text": "@christian Android hardware is less diverse than x86 Hardware, which might have made it easier for Skype to provide a stable service.
", "time": "2022-11-07T16:47:58Z"}, {"author": "George Michaelson", "text": "It would be noted POA in the real world are exploited by bad actors to defraud people in their duty of care
", "time": "2022-11-07T16:47:59Z"}, {"author": "George Michaelson", "text": "not that fraud doesnt happen, nor is not detected, but the model includes bad actors exploiting power
", "time": "2022-11-07T16:48:22Z"}, {"author": "George Michaelson", "text": "loss of agency risks in not understanding delegated powers
", "time": "2022-11-07T16:48:46Z"}, {"author": "David Oran", "text": "Also: this PoA model doesn't seem to accommodate k-out-of-N style authorization where you don't have to put the trust in a single entity that can be hacked.
", "time": "2022-11-07T16:48:54Z"}, {"author": "George Michaelson", "text": "I like M-of-N
", "time": "2022-11-07T16:49:04Z"}, {"author": "David Oran", "text": "Meta comment: please ask the presenter(s) to look at the questions/commentsin the chat. There's lots of good stuff here
", "time": "2022-11-07T16:50:43Z"}, {"author": "Dirk Kutscher", "text": "ack
", "time": "2022-11-07T16:50:52Z"}, {"author": "Wolfgang Beck", "text": "The curse of public key cryptography. Try to explain all this to a non-technical person trying to sign up to some service.
", "time": "2022-11-07T16:51:16Z"}, {"author": "Andrew Campling", "text": "@Wolfgang isn't that a UX question?
", "time": "2022-11-07T16:52:11Z"}, {"author": "George Michaelson", "text": "\"if you ignore this popup and click through you have delegated the right to spend your entire bank balence for 20 minutes do you agree [yes] [no]
", "time": "2022-11-07T16:52:25Z"}, {"author": "David Oran", "text": "Wow: that next-to-last slide looked almost identical to the Kerberos protocol exchanges!!!
", "time": "2022-11-07T16:52:32Z"}, {"author": "Jacob Ward", "text": "Is there any mechanism for revoking a delegation? Does this draft describe a mechanism for invoking such a delegation?
", "time": "2022-11-07T16:53:26Z"}, {"author": "Wolfgang Beck", "text": "@Andrew There not many examples for successful public key UX for the general public. 'make sure there is a green lock next to the URL''h
", "time": "2022-11-07T16:56:39Z"}, {"author": "David Oran", "text": "The short-timeout PoAs are an extremely poor match for IoT onboarding, since devices get provisioned at the factory, shipped and then can sit in a closet for 2 years until needed as a replacement for a failed device.
", "time": "2022-11-07T16:57:14Z"}, {"author": "George Michaelson", "text": "OCSP .. because CRLs dont scale except OCSP doesnt scale...
", "time": "2022-11-07T16:59:21Z"}, {"author": "Wolfgang Beck", "text": "@george true
", "time": "2022-11-07T16:59:50Z"}, {"author": "George Michaelson", "text": "Lucy Lynch did good work on scraping T&C documents off web for analysis. its complicated
", "time": "2022-11-07T17:00:54Z"}, {"author": "George Michaelson", "text": "(for ISOC, a decade back?)
", "time": "2022-11-07T17:01:03Z"}, {"author": "George Michaelson", "text": "(felt like service profiles are in a similar space)
", "time": "2022-11-07T17:01:24Z"}, {"author": "George Michaelson", "text": "the continued implicit emergence of CCITT standards in our life (X509, X.500 as LDAP, nameforms...)
", "time": "2022-11-07T17:07:42Z"}, {"author": "George Michaelson", "text": "if you want to talk centralisation, a globally contiguous namespace in /C/SN/ORG/ terms...
", "time": "2022-11-07T17:08:32Z"}, {"author": "Jacob Ward", "text": "(too late to edit my previous comment, but back on the PoA presentation) How much does this differ from ocap specs such as zcap or ucan?
\nhttps://w3c-ccg.github.io/zcap-spec/
\nhttps://github.com/ucan-wg/spec
no block chain name system today?
", "time": "2022-11-07T17:09:23Z"}, {"author": "George Michaelson", "text": "@Wolfgang hoooray!
", "time": "2022-11-07T17:11:56Z"}, {"author": "George Michaelson", "text": "I like to think an RG sticks to problem statements and measurement/research
", "time": "2022-11-07T17:12:33Z"}, {"author": "George Michaelson", "text": "becoming the 'centralisation directorate' role in a consideration section feels \"out of scope\"
", "time": "2022-11-07T17:12:47Z"}, {"author": "George Michaelson", "text": "(thats not aimed at wolfgang btw)
", "time": "2022-11-07T17:12:58Z"}, {"author": "George Michaelson", "text": "even technical proposals carries risks
", "time": "2022-11-07T17:13:24Z"}, {"author": "George Michaelson", "text": "surely specific charter WG would do tech work?
", "time": "2022-11-07T17:13:39Z"}, {"author": "Wolfgang Beck", "text": "develop technologies that dont scale. If web server could only handle a few clients, the web would be a pretty decentralized affair.
", "time": "2022-11-07T17:18:51Z"}, {"author": "Britta Hale", "text": "These two strands seem to extend scope rather far for one WG. Perhaps it is worth breaking into respective focus groups. Especially the experimental solutions takes on a very different set of problems than the other items listed.
", "time": "2022-11-07T17:19:26Z"}, {"author": "Britta Hale", "text": "They are related, but one is 'big picture' while the other will get into some very technical decision / comparison points.
", "time": "2022-11-07T17:20:35Z"}, {"author": "David Oran", "text": "@wolfgang: in the past those have lost out to ones that do scale, especially in the Internet environment where the usage is decoupled from purely local communication.
", "time": "2022-11-07T17:20:47Z"}, {"author": "Wolfgang Beck", "text": "Current example: today's Mastodon/Fediverse servers dont scale well, so we're seeing a lot of them.
", "time": "2022-11-07T17:21:45Z"}, {"author": "Jim Reid", "text": "\"surely specific charter WG would do tech work?\" Indeed George. Though IMO it's not yet clear what tech work would be appropriate.
", "time": "2022-11-07T17:21:49Z"}, {"author": "Wolfgang Beck", "text": "@david ease of scale and connectivity is a strong driver for centralization
", "time": "2022-11-07T17:26:34Z"}, {"author": "Wolfgang Beck", "text": "centralized: the degree distribution of nodes in the network graph follows a power law
", "time": "2022-11-07T17:30:04Z"}, {"author": "Mallory Knodel", "text": "I\u2019d put a finer point on my statement: I don\u2019t know that the expertise needed to discuss and research internet consolidation is currently present in a group that is chartered to research blockchain.
", "time": "2022-11-07T17:31:09Z"}, {"author": "Wolfgang Beck", "text": "I'd also add network theoretical aspects
", "time": "2022-11-07T17:31:52Z"}, {"author": "David Oran", "text": "@wolfgang - yes absolutely, which is why we need a strong counter-weight. We've been here before, long before the internet, in many industries going back 100+ years.
", "time": "2022-11-07T17:32:05Z"}, {"author": "Wolfgang Beck", "text": "@david yes, free trade had the same consequences
", "time": "2022-11-07T17:33:58Z"}]