IETF 55: Geopriv Requirements Issues

From: Ricardo Cuellar ^lt;JorgeRicardo@web.de>
Date: Mon Nov 18 2002 - 15:48:27 EST

All,

This is the current list of issues identified in the draft
<draft-ietf-geopriv-reqs-01.txt>.

The goal of this list is to identify the changes required to further
advance this document. Please submit comments to the presented issues,
or send more issues to the mailing list. Please refer to the text in
the draft that relates to the issue. If possible, supply also the
suggested text that you would like to see.

Regards,

Jorge

################################################

Current Issues:
Nov 18, 2002

Issue 8b: "Accuracy flag": Closed

  It is not useful to provide an accuracy attribute in
  object, i.e., a flag saying "I'm not telling you the whole
  truth."

  But: if the LO is used for requesting a position, an
  accuracy level may be requested.

  This is an open *protocol issue*: outside of the scope of
  this document.

Issue 13: "LO fields": a) Closed

  What are the contents (fields/attributes) of the LO?
  (This is a "MUST implement", not all location objects
  have to contain all fields)

  1 Target Identifier
  2 Location Recipient Identity
     May be multicast or group identity
  3 Location Recipient Credential
  4 Proof-of-Possession of the Credential
  5 (One or more*) Location Field(s) each containing one or
     more Location Representations, which can be in different
     formats.

    *Issue 15: Out of scope: For privacy reasons, there is
    no need for multiple locations

  6 Location Data Type
     Formats for both lat/long/altitude and "civil" locations

  7 Motion and direction vectors

  8 Timing information:
  a When was the LI accurate? (sighting time)
  b Until when considered current? TTL (Time-to-live)

  9 Policy Field (See also Issue 17)

  MAY be a pointer, e.g. an URI, to a full policy
  or it MAY contain a Limited Policy
  or both

  10 Security-headers and -trailers, e.g. encryption
  information, hashes, or signatures

  11 Version number

Issue 17: "(Limited) Policy Language" = "Core Set": Closed

  Do we want to define a simple policy language?

  Yes, but it may be very simple, say "delete-by date"

  In the Draft: replace the text:
  "The LO should be able to carry a limited but core set of
  privacy policies. This core set is defined below and
  discussed more extensively in a separate document. Beyond
  the core set of privacy policies, the user or Rule Maker
  should be able to define a more robust and complex set of
  policies. "

  By:
  "The LO should be able to carry a limited but core set of
  privacy policies. The exact form or expressiveness of
  policies in the core set or in the full set is not further
  discussed in this paper, but is discussed more extensively
  in a separate document."

Issue 15: " Multiple locations issue": Out of scope

  Is it necessary for the geopriv object(s) to be able to
  carry multiple locations for the target?

  Out of scope: For privacy reasons, there is no need for
  multiple locations

Issue 16: "Full integrity issue": Closed

  Is there a provision in the protocol to prohibit the users
  to send false location information?

  SHOULD the protocol support transformations that introduce
  errors?

  Both: No. There are no such requirements. For more
  discussion, see
  "draft-morris-geopriv-location-object-issues-00.txt"

Issue 18, 19: "Generic policies" (used by LoSi)

  Location Sighters (LoSi) and Ultimate Location Recipient
  (ULR) need in general no full rule-maker defined policies?

  Req. 7. (LoSi Policies) Even if a Location Sighter is
  unaware of and lacks access to the full Privacy Policies
  defined by the Rule Maker, the Location Sighter MUST
  transmit Location Information in compliance with
  instructions set by the Rule Maker. Such compliance MAY be
  accomplished by the Location Sighter transmitting LI only
  to a URI designated by the Rule Maker.

Issue 18, 19: "Generic policies" (used by ULR)

  Location Sighters (LoSi) and Ultimate Location Recipient
  (ULR) need in general no full rule-maker defined policies?

  Req. 8. (ULR Policies) An Ultimate Location Recipient does
  not need to be aware of the full policies defined by the
  Rule Maker (because an ULR SHOULD NOT retransmit Location
  Information), and thus an ULR SHOULD receive only the
  subset of Privacy Policies necessary for the ULR to handle
  the LI in compliance with the full Privacy Policies (such
  as, for example, an instruction on the time period for
  which then LI can be retained).

Issue 20: "Who defines the Identities (Identity Management Issue)"
  Out of scope, see draft Section 9.7

  May the using protocol define the Identifiers or must the
  using protocol use and authenticate Pseudonyms proposed by
  the policies, chosen independently of the using protocol?
  Of course, if the using protocol has an appropriate
  namespace, containing many unused names that may be used as
  pseudonyms and may be replaced by new ones regularly, then
  the Location Object may be able to use the name space. For
  this purpose, the user would probably have to write his
  policies using this name space. Note that it is necessary
  to change the used pseudonyms regularly, because
  identifying the user behind an unlinked pseudonym can be
  very simple.

Issues 21, 22: "Group or role identifiers": Out of scope

  Will the protocol support Role identifiers (like
  "administrator", "member-of-club-A", etc.) Also with
  context dependent meaning?

  Identities may be used to represent groups or multicast,
  but this is outside of our scope. Nevertheless, group or
  role identifiers are probably not somehow explicitly
  supported (at least in V1)

Issue 23: "Disallow anonymous location-requests": Closed

  Need requirement:
  if location recipients decline revealing their identity,
  ==> this must be a designated type of identity,
  allowing the policy to prohibit anonymous location-getting.

  This is not a requirement

Issue 24 b): "Law enforcement issue": Closed

  Law enforcement policies are not under user-control

Issue 25: "Emergency Call Authentication": a) Closed

  Req. 14.3) The protocol SHOULD allow a bypass if
  authentication fails in an emergency call.

  The issue addressed in the last point is that an emergency
  call in some very unfavorable situations my not be
  completed if the minimal authentication fails. This is
  probably not what the user would like to see. The user may
  prefer an unauthenticated call to an unauthenticated
  emergency server than no call completion at all, even at
  the risk that he is talking to an attacker or that his
  information is not secured.

Issue 25: "Emergency Call Authentication": b) Open

  Making an emergency call on a VoIP phone that is not
  "logged in". In the mobile world, you have to be able to
  make a call to emergency services even if the phone is not
  authorized (i.e. it has no service agreement). Problem: if
  the user may not authenticate itself, whose policy to use?
  A default one: this loc information can only be sent to an
  emergency center. Problem: it also the emergency center is
  not authenticable?

Issue 26: " Security Features": closed

  Req. 13. The LO MUST support fields suitable for protecting
  the Object to provide the following security features:

  Mutual end-point authentication
  Data object integrity
  Data object confidentiality

  Replay protection

Issue 27 ("Single Packet Exchange"): Out of scope

  Tracking a small object has several implications:
  1. small device
  2. delta format
  3. The "geopriv protocol" needs to be at most a single
     packet exchange. The first transaction in a tracking
     application could be more than this, but we need a low
     overhead mechanism for incremental update

  Only 2 is now a requirement, but all should be possible.

Issue 28 (Multicast Issue): Closed
  Include the location object in multicast-based using
  protocols.

  Yes

Issue 29 (new): "Full Policy": Out of scope

  Do we want to define a full policy language?
  Perhaps.

  Note: It is outside of our scope how Privacy Policies are
  managed, how a Location Server has access to the Privacy
  Policies, and if he is or not aware of the full set of
  rules desired by the Rule-Maker. Note that it might be that
  some rules contain private information not intended for
  untrusted parties.

Issue 14: "Implementation of a Location Format": Closed

  Location Data Formats?

  The embedded protocol MUST define data type format for
  location data that must be supported by all geopriv
  receivers.

  But not all geopriv LOs have to contain data in this
  format.

Key questions:

  Open issues that are not *protocol* open issues
  Any requirements missing from draft?
  Ready for WGLC?

______________________________________________________________________________
Nutzen Sie Ihre Zeit fur wirklich wichtige Dinge! Alle E-Mail-Adressen
zentral verwalten unter: http://freemail.web.de/features/?mc=021121
Received on Mon Nov 18 15:48:33 2002

This archive was generated by hypermail 2.1.8 : Thu Jan 22 2004 - 12:32:23 EST