Roy Schestowitz <newsgroups@xxxxxxxxxxxxxxx> espoused:
> Government documents suggest Tories not nervous about ISPs interfering with
> Net
>
> ,----[ Quote ]
>| Internal documents suggest the Tory government is reluctant to
>| impose consumer safeguards for the web because it wants to
>| protect the competitive position of businesses that offer
>| Internet access.
>|
>| [...]
>|
>| Net neutrality, dubbed the First Amendment of the Internet
>| in the United States, aims to ensure the public can view
>| the smallest blogs just as quickly and easily as the
>| largest corporate websites. It stops telecom giants from
>| ensuring that pages of companies that pay them load
>| faster than any others.
> `----
>
> http://www.cbc.ca/cp/business/070206/b0206149A.html
Presumably these are the Tories in Canada, although I would expect that
their basic agenda would be much the same as the Tories here.
I do tend to think that the net neutrality debate is something of a
storm in a teacup, since at the transport layer, the network has no idea
where the endpoints are of connection-oriented or connectionless packet
link (consider that for a connectionless packet link, each packet can be
considered to be a single connection-oriented link in its own right, but
has no awareness of previous or subsequent packets).
In terms of actual bandwidths available, this is all the telephony
concept known as Grade Of Service - GOS. Most telecoms networks were
designed to run at a GoS of about 99.9%, ie., 1 call in 1,000 would fail
during the busy-hour. This GoS is achieved by dimensioning the
transport layer for sufficient erlangs, and the switch(es) for a
sufficient BHCA - busy-hour call-attempts.
The concepts are essentially the same for packet networks, the only
present difference being that IP is a lossy protocol, so in IP networks,
unlike in TDM voice-telephony networks, different streams can interfere
with one-another, whereas in TDM networks, they are prevented from doing
so by the network. As carrier-grade ethernet continues to grow in
usage, it will be possible to engineer reliable/predictable packet
network alongside best-effort networking, with either the customer or
the network engineers able to determine which traffic goes down which
pipe. As the complexity of each is much the same, there will be no
reason to charge a different amount for either type of transport, so the
likely result is that customers will, in the end, pay for packets-used
as well as a standing charge for their local loop. This is much the
same as we would currently pay for electricity, say.
As the network providers will therefore charge the same per packet, I
don't really see how net neutrality can be avoided, so long as there is
competition in ISP and network provision. For physical network
provision, where natural monopolies are unavoidable (just watch the US
telecoms market migrating back to one big operator), then regulation
could be required to ensure that the transport businesses do not
interfere with eg., hosting businesses, but this is a rather different
issue to net neutrality in the presently understood sense.
--
| Mark Kent -- mark at ellandroad dot demon dot co dot uk |
| Cola faq: http://www.faqs.org/faqs/linux/advocacy/faq-and-primer/ |
| Cola trolls: http://colatrolls.blogspot.com/ |
|
|
