[AusNOG] ACCC pushes for consumer internet speed test, telcos aren't keen on the idea

[Mark Smith]

On 13 September 2015 at 09:42, Bevan Slattery <bevan at slattery.net.au> wrote:
> The ACCC has created the mess.  They are stupid to push for this as it will highlight their gross incompetence and/or negligence in removing all non-NBN competitive last mile infrastructure (Telstra HFC, Optus HFC etc.) from existing providers.
>

I think ISPs are at fault too. They knew what the CVC charges were,
and could have calculated what their other increased costs would be,
yet they still blindly have tried to provide 100/40, 50/20 etc.
services at prices similar to or low multiples of ADSL ones, as though
the costs of providing them were nearly the same.


> It will more than likely show that TPG's FttB/N network will be superior to every provider on the NBN because they are not economically or technically constrained by either exorbitant backhaul costs to build to 126 PoI's and more specifically the $17.50 CVC charge.  How the ACCC approved this charge as part of the NBN approval has burdened almost every consumer in Australia with the most expensive broadband in the developed world.  I say almost because those lucky enough to have TPG's FttN network will more than likely have excellent value broadband.
>

I think many more POIs was necessary to achieve the necessary
availability, and I wouldn't be surprised to find out that it was the
Defence Department behind it, because in my opinion communications
services availability on a large scale becomes a national security
issue.

To justify its existence (and in theory make a profit), the NBN
network needs to be used for as many services as possible -
residential and commercial Internet, EFTPOS/ATM, Mobile network
backhaul etc. As all of these services are provided externally to the
NBN, the POIs become points of failure for these services.

In the 14 POI model, a POI failure (accidental or intentional
destruction due to fire or other causes I'm not going to mention,
power outage etc.) would impact literally millions of services and
their users. To have these services fail over relatively seamlessly to
the remaining POI would mean having huge amounts of extra resources
(close to if not 100% extra) sitting idle in the other nearby POI just
for that very rare event - and if recovering the failed POI takes any
significant amount of time, then there is now a single point of
failure for all NBN carried communications services in a city (or two
for the Adelaide POIs, as they were also to serve Darwin).

To make this example more vivid, imagine how much worse September 11
would have been if those two buildings had also been the POIs for all
communications services for the state of New York.

121 POIs makes the consequences of a POI failure far less significant.
Switching 100 000 services from 1 failed POI to another near by POI is
going to be a lot easier to achieve than switching 2 million plus
services from one POI to another. With the smaller POIs, you can start
sharing your redundancy resources between groups of POIs, reducing the
amount of redundancy resources you need at each of the POIs. The
problems of performance and availability are easier to solve because
you're allowing the solutions to them to be scaled horizontally.

It also allows more localised services and traffic - in the old 14 POI
model, the Adelaide POIs were also the Darwin POIs, meaning that it
wouldn't have been possible to put CDN nodes in Darwin. The costs of
sending individual, per-user video streams e.g., Netflix and similar
traffic, from CDN nodes in Adelaide to Darwin would have been colossal
compared to serving them from Darwin located CDN nodes.

Cost is an important factor, but it needs to be balanced against the
service availability requirements, and I think the service
availability requirements become much higher when we're moving to "one
network to rule them all", carrying all services.

More diversity usually results in more resilience and performance.
With many networks being operated by many organisations using many
vendors being delivered to many different locations, we've had a lot
of inherent resilience and performance in Internet and other services.
We're losing a lot of that with the NBN, and I'm not completely sure
that has been widely realised.

Regards,
Mark.


> The ACCC have completely undermined their integrity and 20+ years of facilities based competition in this country and are now seen by many who I respect in this industry as a sock puppet for the Minister of the day.
>
> Disgraceful and incompetent.
>
> [b]
>
>> On 13 Sep 2015, at 6:40 am, ausftth at mail.com wrote:
>>
>> "But Mr Sims added the program would cost "low single-digit millions of dollars" with funding coming from either the government or telecommunications providers."
>>
>> Seriously?!
>>
>> The whole thing is a pipedream anyways with CVC charges at $17.50 per mbps. It's a complete waste of money, regardless if millions are spent or not. If ACCC wants to know which circuits are congested they only need ask NBNco.
>>
>> Source: http://www.smh.com.au/business/accc-pushes-for-consumer-internet-speed-test-telcos-arent-keen-on-the-idea-20150911-gjkakd.html
>>
>> Jared
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> AusNOG mailing list
>> AusNOG at lists.ausnog.net
>> http://lists.ausnog.net/mailman/listinfo/ausnog
> _______________________________________________
> AusNOG mailing list
> AusNOG at lists.ausnog.net
> http://lists.ausnog.net/mailman/listinfo/ausnog