Project

General

Profile

Feature #4402

Specification: Create draft

Added by osmith 4 months ago. Updated about 2 months ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
Specification
Target version:
-
Start date:
02/18/2020
Due date:
% Done:

100%


Description

Before we implement this in Osmocom, we will create a draft specification.

Based on the approach described in #4400, I've started a specification document in imsi-pseudo.git:

https://gerrit.osmocom.org/plugins/gitiles/imsi-pseudo/+/refs/heads/master/docs/

imsi-pseudo-spec.pdf View imsi-pseudo-spec.pdf 267 KB draft 2020-04-09 osmith, 04/09/2020 01:21 PM

History

#1 Updated by osmith 4 months ago

  • Status changed from In Progress to Stalled

#2 Updated by osmith 2 months ago

  • Status changed from Stalled to In Progress
  • % Done changed from 10 to 90

As discussed in the weekly telcos, the draft specification does not need to contain the exact references to ETSI specifications that would need to be changed (unlike I had assumed when creating this issue). So the following document should be enough as draft specification, it describes in detail how the IMSI pseudonymization will work:

https://gerrit.osmocom.org/plugins/gitiles/imsi-pseudo/+/refs/heads/master/README.md

laforge: when you have time, can you read through it and confirm that this issue can be closed (that we have completed that milestone)?

#3 Updated by laforge 2 months ago

  • % Done changed from 90 to 50

osmith wrote:

https://gerrit.osmocom.org/plugins/gitiles/imsi-pseudo/+/refs/heads/master/README.md

laforge: when you have time, can you read through it and confirm that this issue can be closed (that we have completed that milestone)?

I think this document currently only works for a reader who has a very detailed understanding of what this project is all about before reading it.

I would suggest to
  • ensure that the document calls itself a specification
  • include a specification version, author, date.
  • include a status (DRAFT)
  • start at a higher level, such as describing how the procedures work in the normal network
  • also describe why the transmission of the IMSI over the radio interface (or to the visited operator) is a problem at all (separately for each case)
  • include ladder diagrams or other illustrations to show the flow of events

Particularly once you start, I think this really should become an asciidoc document. It also provides the infrastructure for versions/authors/status/...

Furthermore, it has the advantage that we could later merge [parts of] it to the osmo-hlr manual.

Would be good to work on this soon-ish, as we cannot complete our first milestone (and invoice) without it.

#4 Updated by osmith about 2 months ago

I've created the proper specification draft with asciidoc in the git repository, and attached the PDF to this issue.

laforge: please have another look.

#5 Updated by osmith about 2 months ago

  • Description updated (diff)

#6 Updated by laforge about 2 months ago

Thanks, looks good in general.

I've pushed some minor changes, please review if they seem OK to you.

Some open topics, from my point of view:
  • we should include reference to the SUCI mechanism in 5G as part of the introduction,
    and point out that this specification wants to provide a similar concealment
    of the permanent subscriber identity - but for legacy 2G/3G/4G
  • we shouldn't add GFDL license to the spec. It's not a user manual. I would consider
    CC-BY for now.
  • We should specify that the mechanism works both for classic SIM but also USIM
  • We should make it less 2G centric. For example, not just the Kc but also the
    keys of other technologies should be removed (EF.KEYS).
  • We should mention that also the P-TMSI (EF.PS_LOCI) and packet switched keys (EF.KEYS_PS) should
    be removed, if present.

Let's address those stated above even in the first draft.

Later on, we should also think of:

  • consider which parts we consider mandatory (the general principle) and which
    optional (the exact binary format of the SMS, which is just an implementation detail and
    doesn't really matter to achieving the desired functionality)
  • consider how IMSI-pseudo would interact within a combined (2G-4G)+5G network,
    whereas the latter already solves the problem by means of the SUCI.

#7 Updated by osmith about 2 months ago

  • Subject changed from Create draft specification to Specification: Create draft

laforge wrote:

I've pushed some minor changes, please review if they seem OK to you.

Yes, they are fine.

Some open topics, from my point of view:
[...]

All changes applied, please review again.

Later on, we should also think of:

  • consider which parts we consider mandatory (the general principle) and which
    optional (the exact binary format of the SMS, which is just an implementation detail and
    doesn't really matter to achieving the desired functionality)

=> #4489

  • consider how IMSI-pseudo would interact within a combined (2G-4G)+5G network,
    whereas the latter already solves the problem by means of the SUCI.

=> #4490

#8 Updated by laforge about 2 months ago

  • Category set to Specification

#9 Updated by osmith about 2 months ago

  • Status changed from In Progress to Resolved
  • % Done changed from 90 to 100

Finished draft pdf is linked in the wiki now: https://osmocom.org/projects/imsi-pseudo/wiki

Also available in: Atom PDF

Add picture from clipboard (Maximum size: 48.8 MB)