Minutes IETF118: satp: Thu 12:00
minutes-118-satp-202311091200-00

IETF Secure Asset Transfer Protocol (satp) WG

Date: Thursday -- 2023-11-09 -- 12:00 - 13:30 Local
Room: Ballroom
25-30ish participants, many online only

AGENDA:

Chair Introduction (Wes Hardaker and Claire Facer)

• Note takers needed: Wes will take notes
• IETF process reminders

WG review of SATP architecture draft (Thomas Hardjono)

• move toward last call
• chairs will spend two weeks reviewing document
• then will hopefully issue last call
• use of github for discussion is ok, just leave issues open for a
  while to encourage discussion

Crash Recovery draft (Rafael Belchior)

• draft ported to markdown
• work ongoing for 3 years about how to recover from crashes
• sort of out of scope at this point but many organizations are
  looking forward to implementing these technologies

Rust implementation of SATP (Rafael Belchior)

• 6 months of work
• work at: https://github.com/hyperledger/cacti
• "relay" block in the implementation implements the SATP draft
• architecture is built on micro services
• 2 relays use SATP between them for collaboration
• currently in pull request 2748
• questions :
  • Chunchi (Peter) Liu: will this handle different DLT versions
    properly?
  • yes the module within the relay can be replaced with different
    drivers to deal with different versions
  • follow on: so the platform SDK needs to be engineered per
    network type?
  • Yes, but everything is agile to being updgraded easily etc
  • DLTs need to operato in their own way
  • Question from claire: is there plans to extend to other networks
  • Yes, but we need people to work on them
  • Mike McBride: it doesn't need to be a DLT right?
  • correct
  • Zainan (Victor) Zhou: can you indicate when an implementation
    implements things where the public records will exist
  • in SATP right now, the relays goes through each network's
    consensus process
  • Wes as chair notes: SATP is designed to work between pulic and
    private networks too, where public verification may or may not
    be possible -- it's a generic transfer protocol

Demo of NFT for DNS ownership use case – (Victor Zhou Zainan)

• Video demo of a SATP like protocol to handle DNS transfers between
  registrars and registries
• demo of Alice transfering a domain to Becca
• need to coordinate locking/minting
• need to coordinate payment involved

Report from the Network Identification Design Team (Thomas Hardjono & Weija Zhang)

• three general goals:

  • need to deal with cases where there may be multiple forks in a
    network
  • internally self-identification is important
  • network ownership is a legal issue

• proposal uses a TLV

• detailed description of their current proposal (see slides)

• questions:

  • denis: it's impossible to understand in L2 which DLTs you are
    anchoring, so there is a whole discussion about how to identify
    L2 networks

• chair notes:

• time needed to complete your proposal?

  • answer: likely by march

• reminder: the design team merely gives a recommendation, wg gets the
  final call on the solution,

• type for private usage identifier or range between understanding
  networks, typically at the end
• do the datatypes your proposing properly match with SATP core
  protocol

Asset profile definition (Denis Avrilionis)

• NOTE: that this is potential "future work" for the working group,
  not in the active charter
• describe as precisely as possible the asset being transfered to help
  whether the transfer could be accepted on the other side
• can the asset be defined in a more generic way
• define a general definition of what an asset scheme is
• examples: DNS domain from the other presentation, bills of lading

• questions:

  • Rama: the network id of the previous network could be used for
    tracabality
  • Denis: yes, but it could be more than the network such as the
    formal instance from the previous network as well in some cases
    since the previous was burnt
  • Rama: suppose a gateway crashes and then recovers but can't log
    an asset because it's crashed, it may not be able to log it
    since another entity has already locked/reserved it
  • Denis; good point, we will need to carefully figure out how to
    do that (with more thoughts not captured in notes including
    pessimistic or optimistic logging mechanism)
  • Peter

• 15m - AOB

• future interim meetings? Should we have them?
• chairs will start a doodle poll on the mailing list to determine
  best times

Related documents: