[Linux-aus] SIP

[David Lyon]

On 03.07.2013 02:55, David Newall wrote:

> Domains are an almost infinite resource and it makes perfect sense to 
> me
> for every person to have a few of their own.  That doesn't mean that
> they must provide their own infrastructure, a fact which is clear 
> from
> brief examination of web and e-mail addresses.

That was predicted a decade ago and there was a rush to sell domain 
names after all logic says there are 4+billion people on the planet - 
all needing their own domain name.

It didn't pan out that way, and there were some business failures in 
the domain-name selling industry.

Personally, I wouldn't see domain-names as that beneficial to people 
with respect to the nbn. Having said that, dns server's with 20 million 
records are not that big a technical burden these days. That's just one 
rack or less of blade servers - correct me if I'm wrong.

> A practical way to enter the VoIP game using non-proprietary 
> protocols
> is to take one of the many find consumer-grade modems or routers and
> replace the firmware with a community-supported alternative such as
> OpenWRT, for which Asterisk is available.

That's an interesting idea, but that's not the main issue.

The heart of the issue is that the core business of any 
telecommunications company is usually defined as 'Billing'.

Without 'Billing', telecommunications usually doesn't work. That's the 
simplistic explanation of why SIP hasn't been successful. It has no 
associated Billing System.

> But, again, may I say that the technology to run voice, video, chat 
> and
> shared-desktop over internet is already well mature even as it 
> evolves.
> What I see holding back universal adoption is the concern over
> interoperability with the legacy system.

Put another way, the income generated by conventional Billing systems 
means that concern is placed over 'free' voice, video, chat and 
shared-desktops.

The Desktop sharing scenario is quite interesting. The main player in 
that game is LogMeIn. To the best of my understanding, the two problems 
that they have solved are a) Billing and b) DNS.

Apart from that, the program looks like it could well have been built 
on Open-Source components.

I'm unaware that Desktop-Sharing is part of the NBN.

> This is a hard problem which,
> presently, is mostly answered by hiding the new technology behind an
> obsolete number.  Absurdly, even businesses who have bought in to 
> VoIP
> hide the fact.  Try calling IBM using VoIP and you will succeed, but
> only by using their legacy numbers.  IBM hides, or at least declines 
> to
> highlight details of, their VoIP systems.  The world keeps ignoring 
> the
> elephant in the corner, which is that telephone numbers are a very 
> early
> 20th century invention that has outlived its need.

I disagree with that from a telecommunications standpoint.

A telephone "Number" is a convenient Billing-System identity. It's easy 
to print on a Bill and easy to fill with cost-item-entries (Calls/SMSs).

To render a Telephone Number obsolete in the way that you suggest, a 
new way of Billing the customer would need to be introduced.

Until a better Billing-Identifier is presented, I'm sorry, you won't 
have much 'forward-movement' of Billing Systems in the way that you 
would hope.

It's not that these things represent my personal view, but more that 
they form the basis on how the Telecommunications Industry is built, and 
more importantly - Funded.

-David