Minutes of the 2012-07-31 IAB Business Meeting
Vancouver, BC, Canada

1. Attendance

PRESENT

• Bernard Aboba (IAB Chair)
• Jari Arkko
• Mary Barnes (IAB Executive Director)
• Marc Blanchet
• Ross Callon
• Alissa Cooper
• Spencer Dawkins
• Mat Ford (ISOC Liaison)
• Joel Halpern
• David Kessens
• Cindy Morgan (IAB Executive Assistant)
• Jon Peterson
• Dave Thaler
• Hannes Tschofenig

APOLOGIES

• Lars Eggert (IRTF Chair)
• Russ Housley (IETF Chair)
• Danny McPherson
• Robert Sparks (IESG Liaison)

GUESTS

• Dan Romascanu (item 2 only)

2. Followup from the IEEE 802/IETF Meeting

Spencer briefly reviewed the outcomes of the meeting with the IEEE 802 leadership held the previous week. A team is being formed to write an update to RFC 4441, which outlines the relationship between the IETF and the IEEE 802 with participants from both sides.

The IEEE 802 and the IETF leadership also agreed to plan conference calls every two months to coordinate ongoing work, with a plan to meet face-to-face in Orlando around IETF 86.

The board discussed appointing a liaison from the IETF to IEEE 802.11; Dan Romascanu agreed to come up with a plan for the best way to handle this.

3. Liaison with ETSI M/493 for Emergency Services

Alissa reported that the Emergency Services Program met and discussed ETSI M/493 that is working on Emergency Services in Europe. The board discussed whether to send a liaison to this group, and agreed that the liaison would be drafted by the Emergency Services Program, reviewed by the ECRIT Working Group Chairs, and approved and sent by the IAB.

4. Confidentiality in IAB Programs

The board approved the text Spencer proposed for the message on confidentiality in IAB Programs. The Program Leads will forward this message to their Programs, and the text will be posted on the IAB website.

5. Affirmation of the Modern Global Standards Paradigm

Bernard reported that IEEE-SA and the W3C have proposed edits to the “Affirmation of the Modern Global Standards Paradigm” document. The IAB will discuss this further with the IESG on Friday.

6. ICANN BOD Position

Jari reported that Thomas Narten is willing to continue as liaison to the ICANN Board of Trustees until a successor can be found. The board agreed to issue a call for nominations in January and plans to select a new liaison by Spring 2013. Thomas will be asked to provide input into the ICANN Board of Trustees Liaison job description.

7. Technical Topic: The New Hourglass

Marc and Hannes led a discussion on HTTP/HTTPS as the new waist of the TCP/IP protocol stack. Following are the slides presented during the meeting.

–Begin Slides–

  1. Internet over port 443/New hourglass
     Marc Blanchet
     Hannes Tschofenig
     IAB meeting
     July 31st 2012
     Vancouver

  2. Plan
    - 2 drafts on a related topic, different perspective
    - Presenting each
    - Comments/discussion

  3. Draft-blanchet-iab-internetoverport443 in a nutshell (1/3)
    - Claims that (too) many access networks are only allowing port 
      80/443 outbound connections.
    - Therefore, only way to deploy an app reliably in all cases is to 
      use port 80 (most likely http based) or port 443 (« easier » to 
      carry your own protocol)

  4. Draft-blanchet-iab-internetoverport443 in a nutshell (2/3)
    - Many impacts, some listed in the draft
      * Traffic policying based on the addr+ports is « not possible » as 
        is, since all (very different) trafic goes over the same port
      * Deploying new protocols and new transports is « not possible »
      * Overloading http : « for my new protocol, I need this and that 
        new http header. »
      * Increasing the uneffective use of IP addresses : worst case : 
        two non-http services use two IP addresses on the server side
      * More complex network operations : netflow graphs see only two
        ports...
      * All apps have to be over HTTP/websockets?
      * The Internet transport is HTTP.
      * ...

  5. Draft-blanchet-iab-internetoverport443 in a nutshell (3/3)
    - Possible recommendations
      * (not in the current draft, but in the local copy, awaiting 
        agreement on the overall problem statement) :
        o If you are inventing a new protocol or using a different port
          than 80/443, please read these issues (and justify?)
        o Should look for your « transport » at : « plain » http/
          restful, websockets, …
        o BCP to access providers : please open your ports...
    - Bigger (IAB/IETF wide) question :
      * Except for « infrastructure » (i.e. Provider backbone kinda of), 
        are we in a port 80-443 only world for standards?

  6. The New Waist of the Hourglass
     draft-tschofenig-hourglass-00.txt

  7. The Original IP Hourglass
  +------+------+------+------+------+
  |Email | Web  | VoIP | P2P  | RTSP |
  +------+------+------+------+------+
         | TCP  | UDP  | ICMP |
         +------+------+------+
                | IP   |
         +------+------+------+
         |Ether |Sonet | ATM  |
  +------+------+------+------+------+

  As everyone knows this model had changed due to NATs.

  8. NATs and Firewalls
    - There are, however, other challenges for ICMP and for getting 
      transport protocols other than TCP and UDP deployed.
    - Example consequence: standardization work for UDP over SCTP.
    - Connectivity checks to test e2e reachability and more multiplexing 
      (to reduce the number of tests)
    - IPv6 deployment is not likely to lead to a change.

  9. HTTP
    - With the rise of the Web developers are increasingly familiar with 
      HTTP. They feel comfortable with it and use it as a foundation for 
      lots of protocol designs (voluntarily).
    - While there is filtering applied to networks it is
      not clear that developers have to use HTTP/HTTPS since the waist 
      had moved again.
    - Data points available but far too few.

  10. Possible Next Steps
    - Not an option: Tell people to stop use HTTP/HTTPS
    - Encourage community to investigate the degree of filtering in 
      various network environments. Examples:
      * Open Observatory for Network Interference
      * Neubot project
      * “An experimental study of home gateway characteristics”
      * “An Untold Story of Middleboxes in Cellular Networks”

  11. What's next?
    - IAB adopt one, both, merge,??
    - Discuss/adopt in IPevol first?
    - Comments?

–End Slides–

After discussion, the board agreed to discuss these drafts further in the IP Evolution Program to determine whether there is work to be done.

8. IANA Evolution

Bernard reported that the evolution of the IANA function will be discussed at the I* leaders’ meeting in September.