Regulatory intervention

Erik Haagsman erik at we-dare.net
Fri Oct 7 15:07:15 UTC 2005


On Fri, 2005-10-07 at 14:56 +0100, Michael.Dillon at btradianz.com wrote:
<snip>
> Laws only need to be enforced when there is a dispute.
> Laws and regulations, do not necessarily imply that
> enforcement action is needed. Many people and organizations
> comply with laws for reasons other than the existence of
> enforcers. For instance, an organization may feel that it
> is in the industry's best interests to comply with regulations
> and therefore it does so in order to set an example for
> its competitors and to attract customers.
> 
> Regulations also do not imply the involvement of governments.
> It is possible for industries to self-regulate such as the
> ARIN policies which are a product of the ARIN membership,
> i.e. companies who use IP addresses in their networks.

Very good point and IMHO the preferred way of dealing with these kinds
of issues without the overhead of specific legislation and often
stifling governmental intervention. The approach you outline below seems
very plausible, with a regulatory organisation of some sort driven by
the industry itself protecting both ourselves as well as our customers
from idiocy like the whole Cogent/L(3) thing. It would improve both
better interconnections and network coverage (and thus network quality
IMO) as well as more transparency in peering and interconnection
relations. Both good things for end-users and xSP's alike.

> If the press would truly understand this event then they would
> be reporting this as a *MAJOR* flaw in the business model of 
> the largest ISPs. The absence of regulation in Internet peering
> allows this type of situation to come about. It is my opinion
> that the network and the Internet business would both be stronger
> if there was some regulation of peering and IP/MPLS network 
> interconnection. 
>
> This could be done in a couple of ways. One is to have an industry
> association develop self-regulation in conjunction with major end
> users of network services. The other would be for regulation to be
> imposed from without by some kind of interconnect or monitoring
> business like Equinix or Keynote. The analogy here is the New York
> Stock Exchange which is a 3rd party which monitors and interconnects
> the buyers and sellers of shares. In the case of Internet operators
> I don't foresee the need for an SEC equivalent unless operators
> cannot agree to disclose their peering agreements and the technical
> details of their interconnects.
> 
> A couple of good things can come out of this "open peering" model.
> One is that disclosure of the technical details, including packet
> drop, buffer consumption, and bandwidth, would lead to more reliable
> interconnects and the ability to provide quality of service SLAs 
> across provider networks. The other possible benefit is to develop
> more sophisticated interconnect variants such as MPLS VPN interconnects
> and CDN or multicast interconnects.
> 
> --Michael Dillon
> 
-- 
---
Erik Haagsman
Network Architect
We Dare BV
Tel: +31(0)10-7507008
Fax: +31(0)10-7507005
http://www.we-dare.nl





More information about the NANOG mailing list