Course: CS125
Name: Daniel Meredith
Commentary Due Date: 01-26-00
Submit Date: 01-26-00
Journal Ref: IETF RFC2026 & IAB Charter 1994
The IAB (Internet Architecture Board) is the governing body of the internet,
if such a thing exists. The latest copy of their charter that I was able to
locate was from 1994, but it seems to give a descent idea of what the IAB is
about and how it accomplishes the somewhat daunting feat of organizing and
running the administration aspects of the global internet.
The IAB is composed of 13 members. The IETF (Internet Engineering Task Force)
and 12 other full IAB members. They are all able to vote on all IAB matter,
with the exception of the IETF chair, who is not able to vote on matters
involving the approval of IESG (Internet Engineering Strategy Group) members.
Members of the IAB are nominated from a list provided by the IETF nomination
committee. There are generally six new members nominated every year. The board
members serve terms of two years, but are not limited in the number of inline
terms which they may serve.
The role of the IAB is defined in the charter to be six main responsibilities:
- IESG Selection
- Architectural Oversight
- Standards Process Oversight and Appeal
- RFC Series and IANA
- External Liaison
- Advise the ISOC
The first role is simple enough. The IAB elects the IESG members to their two
year terms. Of the twelve voting members of the IAB eight must approve of the
nomination to the IESG, but their can only be one negative vote. This somewhat
odd to me. There is a possibility of a non-vote by three of the twelve (25%). I
feel that this is inappropriate and a vote or at least give an opinion on the
nominations.
The second role is to give architectural oversight to the standards being set
by the IETF. This conjures the image of the overlords in some dark room
deciding the fate of the internet. But in actuality they simply help guide
the new working groups and check for consistency and integrity in their work.
They also oversee the IRTF (Internet Research Task Force).
Standards which the IAB set are put through a rigorous process which I will go
into more detail about later in this commentary.
The IAB also must manage all the documentation produces by the RFC and BCP
series of standards. They are responsible for the upkeep of the documentation
on all the development done by the IESG and IETF. The IAB must also number and
organize the documentation. The IAB designates the IANA (Internet Assigned
Numbers Authority) to administer the assignment of protocol numbers.
The IAB can also act as an external liaison and advisor to various groups
around the world. In an advisory role the IAB is most involved with the ISOC.
The IAB may convene panels of experts to investigate questions and topics
presented to the ISOC.
Overall the IAB seems to be well organized. But like many organization the
charter and reality can be very different. I have need been to and IAB meeting,
nor have I ever had to interact with them at all in the past. Like most users
of the internet I have just trusted that I would continue to work, similar to
the sun. It is not always up in my area, but it is up somewhere. I see a few
problems in the organization of the IAB.
- The groups is required to come to consensus on issues. This will not work
as the groups begins to grow. There are only thirteen voting members, but few
issues are voted on. Most require the consensus of the group.
- The IAB seems to get its members from the same groups that it appoints.
This seems to be continually pulling from the same pool of people over and
over. Perhaps the IEEE, ISO, and ANSI should be included in the pool of
possible nominees.
- In the spirit of government reform, perhaps some form of term limits should
be imposed on the voting members of the IAB and the IETF chair. This would help
bring new blood into the organization and perhaps help speed along the
development of the internet.
The second document we were asked to read was the RFC that outlined the process
in which new protocols and possible standards become the standards of
communication on the internet. At over 30 pages the document is quite specific
in the definition of how a working group can create a standard. I will not go
into the specifics of how the process works, but instead give an general
overview.
A working group in the IETF publishes their preliminary document in the form of
an Internet Draft. This document only has a life span of six months. If work
is stopped on it during that period it essentially dies. If a protocol makes it
through the Internet Draft stage it is then moved to an RFC (Request for
Comment) in the Proposed Standard stage. The protocol is then edited and
revised until it moves to the Draft Standard stage after about six more months.
At this point the working group must show two independent implementations of
good running code. After another year or so the draft may then move to Internet
Standard, or simply standard. This is the final place for the protocol. It
be revised and re-worked some over the next few years, but it is now a
standard.
That was the quick and dirty of how a standard is created, and it seems very
dry, but the interesting parts are in the appeals process. An appeal can be
launched from any level in the IAB/IETF structure. It then moves up the
structure until it reaches the IAB at which point the decision is final. The
RFC for the standards protocol specifies that there are two main basis for
appeals:
- unfair representation of ideas within a working group
- incorrect technical decisions which place the quality of the product in
jeopardy
These are very different types of appeals, and yet they are handled in the same
manner. The first is the belief by an individual that their ideas or an idea is
being unfairly dismissed by the working group. This is a person-to-person
issue, but also a technical issue. The latter is simply a technical issue:
Was an engineering mistake made somewhere along the line?
Both types of issues require a technical expertise to decide, but the first is
much more complex. This appeals are sent through the same system of appeals.
This is wrong. There should be a separate channel for the first type of
appeal due to the fact that it is very centered around the working atmosphere
of the working group.
The other major oversight I noticed in the appeals process is that the bodies
which decide the appeal at each level are given their own discretion as to the
methods in which the appeal should be decided. There is not set standard. And
in a body which is entirely concerned with setting standards this would seem
odd. By requiring the appeals bodies to follow a set protocol the fairness of
the appeals process.
The charter and the RFC give a good look into the organizations that run the
internet. They are not perfect, but for a group of unpaid individuals they
obviously take their role in the development of this new world very seriously.
The root of the matter is the that the IAB and the groups under it make the
internet possible. By developing with intention of open standards and the free
exchange of knowledge the IAB makes the internet one of the most powerful
tools ever put in the hands of mankind.