Spending a few days at the Internet Identity Workshop 35 in San Francisco provided a great opportunity to catch up on some reading.
I was especially curious about one of the recent items that was published by the Trust Over IP Foundation, the first public review draft of the Technology Architecture V1.0 Specification.
Those familiar with the Trust Over IP are likely familiar with the Dual Stack Architecture Model, a framework that adds a governance layer across the technology side of things. The practical governance and policy questions that must be answered in order to drive business, legal, and social acceptance.
The new Technology Architecture Specification now changes the way we look at what is required to attain digital trust.
The hourglass model is perhaps the most widely recognized hallmark of the Internet protocol suite. The “hourglass” metaphor is evident from looking at a diagram of the TCP/IP stack such as Figure 5 from a presentation by Steve Deering of Cisco given at the IETF 51 meeting in London in August 2001.
The single “spanning layer” at the “waist” of the hourglass is the IPv4 protocol. However, the components that lie above the distinguished layer cannot directly access the services that lie below it. The rest of Mr. Deering’s presentation delves into various efforts to “fatten” the “waist” of the hourglass by “squeezing other protocols into the middle”. He strongly recommends resisting this.
Much of the success of the Internet is attributed to this “hourglass” design in which the spanning layer protocol Maximizes interoperability by providing a common way for all the higher level layers to communicate with all the lower levels. This is why the design of the trust spanning layer should be “as simple as possible but no simpler”.
The new ToIP Technology Architecture V1.0 Specification has applied this same hourglass design to the four ToIP layers.
This new design makes it much simpler to start seeing how a multi-protocol approach can work. One of the more interesting conversations for myself at IIW 35 was about how OpenID4VC will be able to work within this model. I’ve tried to summarize some thoughts in some Tweets and will try to write some more about it in the weeks to come.
https://t.co/wO0QaybPF6— Mathieu Glaude (@mathieu_glaude) November 17, 2022
DIDComm and OpenID4VC are apples and oranges and can both work in conjunction within the new ToIP Technology Architecture proposal.
There is a list of 18 requirements in the specification that relate to the Trust Spanning layer of the hourglass model, and DIDComm is proving to be a good candidate for this layer.
How does the existing ToIP Dual Stack Architecture Model transition towards the new hourglass model?
This is the way I’m looking at it now, but I welcome any feedback.
Just like the Trust over IP welcomes feedback to the new public review draft of the Technology Architecture V1.0 Specification. So please go check it out and provide your thoughts as well.