Review: Tussle in Cyberspace: Defining Tomorrow's Internet

The paper identifies a process the authors call "the tussle" in which
different stakeholder that are part of the Interent milieu have
intersets that may be adverse to each other, and these parties contend
to achieve their conflicting goals. The authors claim that the challenge
facing Interenet research and engineering is to recognize and leverage
this tussle, to at least accomodate it and if possible use it to
strengthen the technical architecture.
The highest-level priniciple to deal with the tussle is to design for
variation in outcome, since rigid designs will be broken and designs
that permist variation will flex under pressure and survive. The authors
emphasize that rather than design the outcome, we should design the
playing field for the tussle. The high-level principle leads to two more
specific principles:
1. Modularize the deign along tussle boundaries, so that they don't
intefere with each other.
2. Design for choice, to permit the different players to express their
preferences. However, they recognize that this will add to complexity of
configuring and using a service.
The paper discusses specific aspects of the Internet tussle through
examples and tries to illustrate how the principles provided can be
applied. The authors discuss examples in the areas of economics (such as
value pricing), trust and the tussle of openness. The authors then
redefine the following classic principles in more complex terms:
- end-to-end argument:the old end-to-end argument called for more
transparency whereas now loss of trust and desire for control by ISPs
calls for less transparency.
- separation of policy and mechanism: true-value neutral design is
difficult. The advantage is to isolate some regions of the system from
tussle and use these to separate different tussles from each other.
The paper discusses an interesting and relevant phenomenon. However, one
of their main principles is to modularize design along tussle
boundaries, which in my opinion will most likely not be clearly defined
and just identifying them is a difficult task. Furthermore, while
talking about competitive wide area access, the authors say that the
Internet should support a mechanism whereby customers can control the
path of their packets at the level of providers. I disagree because I
think allowing eacg customer to choose source routing won't be scalable.
Received on Mon Nov 27 2006 - 17:35:26 EST