Enhancing the Internets Administrative LookUp Service

1
Enhancing the Internets Administrative Look-Up
Service

• Andrew Newton
• Leslie Daigle
• Mark Kosters
• VeriSign Labs
• APRICOT, March 2002

2
UWhat?

• Universal Whois
• VeriSign has committed undertaking in agreement
  with ICANN
• Formal public consultations
• business, intellectual property holders (Aug/01)
• civil liberties, other ngos (Nov/01)
• international input (Nov/01)
• Informal public consultations
• RIPE 40 (Oct/01)
• NANOG 23 (Oct/01)
• RIPE 41 (Jan/02)
• NANOG 24 (Feb/02)
• APRICOT 2002 (Mar/02)

3
Community at a Glance

• If we tried to include every aspect of every type
  of whois service (past or present) in the world,
  we would never get any work completed. The scope
  would be too large.
• The subset is the community of people that
  administer and use the Internet
• Network operators and service providers
• Registry operators
• Implementers of software (for this community)
• Registrars, Certificate Authorities, etc.
• IPR Owners, Law Enforcement, Policy Groups, Other
  Governmental Agencies, Non-Governmental
  Organizations, etc

4
So is Harmony Communal?

• Not always
• Law Enforcement and some Non-Governmental
  Organizations sometimes want conflicting goals.
• Intellectual property and privacy advocates also
  appear to have conflicting goals
• The laws applying to various network and registry
  operators vary from country to country.
• the list goes on
• We must provide the mechanism, not the policy.
• Because it is not our job.
• And we would never finish if we did.

5
Some of the Potential Requirements

• Structured queries and results
• Referrals and referral-path authority
• NIC Handle references
• Standards
• Ease of implementation and minimal re-invention
• Machine readability
• Decentralization and one-stop-shopping.
• Privacy and access by IPR holders and law
  enforcement.
• Adaptable to many policies and laws.
• After 30 years of Internet Science, it can be
  done.

6
Discussion

• When we list out some of the requirements, they
  cause us to ask more questions?
• Your input is needed.

7
Referrals and References

• Referrals are needed to traverse the many
  repositories of information.
• DNS could be referral-path authority when
  applicable (SRV RR).
• What are the perceived issues with making
  referrals DNS based as opposed to an out-of-band
  mechanism?
• What about other hierarchies that dont map to
  DNS?
• How do you reference a handle from service A in
  service B?
• URL (e.g. LDAP URL)?
• URN?
• Ideas?

8
Implementation

• Stay on port 43?
• Everyone knows port 43!
• Backwards compatibility issues
• Dont want to get bogged down in the niche/other
  uses of whois (whois BOFs at IETF 49, 51).
• Dont want to have to re-invent
  security,internationalization, etc features.
• Not 43?
• Fresh start!
• Dont invent too much.
• Does the world really need another application
  transport?
• Does the world really need another schema
  language?
• If we are smart, we can re-use a lot of existing
  building blocks to create a new protocol (e.g.
  BEEP, SASL, XML).
• Possibility of existing protocols LDAP, whois,
  Rwhois, UDDI.

9
(De)Centralization

• Aggregation of all (or even much) of this data
  into a central repository wont work.
• In some countries, there are laws and policies
  that forbid it
• Also some good technical reasons.
• Besides, isnt part of internet-working about
  decentralization?
• A centralized view or one-stop-shopping really
  helps usability.
• Context-free Search Continuation References
  (referrals) solve only part of the problem.
• Good for lookups. Limited for searches.
• Common Indexing Protocol could help searches.
• Referrals are context-sensitive.
• Aggregates the hints, not the data.
• What level of obfuscation is acceptable?

10
Potential Privacy and Law Enforcement

• Privacy concerns vary from jurisdiction to
  jurisdiction
• Define the technical mechanism, not the policy.
• Consider current mechanisms (e.g. TLS, SASL).
• One option when to give access to restricted
  information can be tricky. Potential solutions
• Distribution of authentication tokens.
• Logistically challenging.
• Verification via digital certificates.

11
Conclusion

• Your comments, opinions, and ideas are welcome.
• http//uwho.verisignlabs.com/
• Further reading
• Requirements
• draft-newton-ir-dir-requirements-00.txt (coming
  soon)
• LDAP proposals
• draft-newton-ldap-whois-00.txt
• draft-hall-ldap-whois-00.txt
• XML proposal
• draft-newton-xdap-01.txt
• draft-newton-xdap-domdir-01.txt
• draft-newton-xdap-ipdir-01.txt
• The State of Whois
• draft-campbell-whois-00.txt
• draft-brunner-rfc954-historic-00.txt