WEBVTT

1
00:00:05.800 --> 00:00:10.212
SCIR stands for Self-Certifying Identifier Reference.

2
00:00:11.717 --> 00:00:19.484
It identifies an object and lets us verify the binding between its identifier
and presented certifying material.

3
00:00:21.008 --> 00:00:24.143
The same envelope distinguishes four classes.

4
00:00:25.967 --> 00:00:30.053
IDENTITY references an identity bound to a public key.

5
00:00:32.175 --> 00:00:42.423
ALIAS references canonical text that serves as a rendezvous for interest-based
discovery. Different interests can have their own aliases, a participant can

6
00:00:42.423 --> 00:00:46.720
associate with several, and a query can find multiple candidates.

7
00:00:48.842 --> 00:00:51.895
MULTICAST references a logical transmission.

8
00:00:54.008 --> 00:00:55.843
GROUPCAST references a group.

9
00:00:57.967 --> 00:01:04.666
For multicast and groupcast, the identifier is derived from immutable genesis
material.

10
00:01:06.175 --> 00:01:12.454
The first byte carries the class in its lowest two bits. The remaining six
bits must be zero.

11
00:01:14.383 --> 00:01:20.119
The next two bytes identify the reference suite, most significant byte first.

12
00:01:22.050 --> 00:01:26.659
The identifier follows, with a length defined by that suite.

13
00:01:28.592 --> 00:01:36.057
The complete SCIR contains three header bytes plus the identifier length
defined by the suite.

14
00:01:37.592 --> 00:01:46.497
The current suites use 32 or 48 identifier bytes, producing references of 35
or 51 bytes.

15
00:01:48.008 --> 00:01:51.155
New suites can define other identifier lengths.

16
00:01:52.675 --> 00:01:56.425
The table compares the five current reference suites.

17
00:01:57.967 --> 00:02:07.951
They use Ed25519, ML-DSA, or a combination of both, with the hash and
identifier length shown in each row.

18
00:02:09.467 --> 00:02:17.338
The public-key sizes apply to IDENTITY; the hash and identifier length apply
across the four classes.

19
00:02:20.175 --> 00:02:29.602
Self-certification means that we can recompute the identifier from presented
certifying material and compare the result with the expected reference.

20
00:02:30.925 --> 00:02:36.614
For IDENTITY, we first validate the public key's canonical encoding.

21
00:02:37.925 --> 00:02:42.848
The Strong ID is the hash of a fixed domain, the suite, and that key.

22
00:02:44.175 --> 00:02:51.524
An ALIAS lets us look for participants associated with an interest instead of
selecting an individual identity in advance.

23
00:02:52.842 --> 00:03:01.166
The relationship can be many-to-many: several participants share an interest,
and each participant can have multiple interests.

24
00:03:02.467 --> 00:03:11.801
The application defines what those interests mean; PRP handles the canonical
text used to form each reference.

25
00:03:13.133 --> 00:03:17.336
For an ALIAS, the certifying material is canonical text.

26
00:03:18.842 --> 00:03:23.985
We first apply Unicode NFC normalization and UTF-8 encoding.

27
00:03:25.300 --> 00:03:27.645
The length counts bytes, not letters.

28
00:03:28.967 --> 00:03:33.088
Case, punctuation, and spaces remain significant.

29
00:03:34.425 --> 00:03:40.323
Here we use a specification vector and the suite with SHA-256.

30
00:03:41.633 --> 00:03:48.692
The hash receives a dedicated domain, the class, the suite, the length, and
the canonical text.

31
00:03:50.008 --> 00:03:54.919
Repeating the calculation produces exactly the same reference.

32
00:03:56.258 --> 00:03:59.068
This establishes the binding to that text.

33
00:04:00.383 --> 00:04:06.966
A SCIR contains a class, a reference suite, and a suite-defined identifier.

34
00:04:08.800 --> 00:04:13.990
Self-certification verifies the binding to the presented certifying material.

35
00:04:15.800 --> 00:04:16.845
PRP.