[LINK] Ergas article in the Oz

[Tom Koltai]

It's interesting to note that these discussions are appearing during the
time of consideration for the ACCC in reference to the Foxtel
notification of delivery via the Bigpond service.

Form G Notification of Exclusive Dealing Conduct [N94674]


Whilst I didn't get my submission in to the ACCC in time, I have decided
to blog about it.

In a series of three articles I will explain how it is that Telstra can
deliver the bandwidth on an anti competitive basis, where no other ISP
can "legally" go. 
In fact, Telstra are using P2P to deliver their new VOD services.

The first of the three articles is at http://kovtr.com/wordpress/?p=910

And reads:

This article is Part one of a three part series and presents the case
that even P2P can become legitimate if the right organisation starts to
use it.

Preface:

The following article is a much simplified technical discussion about
the Australian HFC networks delivery of content.

Certain liberties for ease of understanding have been taken by the
author.  For example, I quote 5 Mbps bandwidth on cable, when most
people have seen much higher speeds.  This is justified with the
explanation that the number of users on a cable segment and their
interactive usage habits contribute greatly to the amount of available
bandwidth (download speed) available at any one time to any one user.

----------------------------------

Many of the articles penned here appear to knock the denizens of our
content and legislative peers.

Yet, if no-one ever complains about anything, then how will anything
ever improve.

Today's article will at first appear as if I have lost my mind. [For
those convinced that I have already achieved that milestone, I need not
deliberate any further, for those whom consider my humble pearls of
wisdom of interest, I regret I am about to apparently disappoint.]

This article is the first article that I have penned that lauds the
technical advances that Telstra have made in content delivery.

Sometimes revolutions are noisy affairs with lots of publicity, shooting
of ministers, generals and other perceived enemies of the state.

Occasionally, the new order is announced, installed and hardly anyone
notices.

Last month a new global technology revolution was announced in Sydney,
and it was greeted with nary a whimper.

Yet the technology that Telstra announced for the delivery of their
Video on Demand content stream has been the contention of industry vs
the
consumer since Napster was announced.

Yet as usual I get ahead of myself.

Each new technology disrupts the forerunner.

Table 1 Transportation Methods
[table]

Each "upgrade" created more traffic and paths turned to lanes, roads,
highways and eventually motorways.

.and more traffic.

Table 2 Communications
[table]

With communications, each level of upgrade increased consumer knowledge,
participatory commerce and consequently, individual nations economies.
i.e.: the countries with the liveliest and lowest cost communications
infrastructure became the highest standard of living "industrialized
nations".

Quite often the elements of each of the above tables are utilized
together to deliver an item of communication to the recipient.

For example, A telegram circa 1910 would have been delivered via a
bicycle, but by 1920, via motorcycle.

Did the telegram delivery persons get upset ? Of course not, they had
been given shiny new motorcycles.

The telephone is an interesting analogy for the lack of resources that
the internet is about to experience. [I'll explain that statement
shortly.]

In the 1930's, it was common to deliver a single pair of copper wires to
a street and then run extension lines from that first installation.

In this manner, connectivity was achieved a lot quicker than running
separate copper pairs from the exchange to each individual house.

Technically it was feasible, but the cost would have been prohibitive.

This system of sharing a single copper between multiple houses was
called "party-line".

One ring was for house number 1, two rings was for house number 2 and
three rings for house number three (etc). In some areas, there were
15-20
houses on a single circuit, in which case there was a code implemented,
one short and two long was for house 15, two short and one long was for
house 11 etc.

It was clumsy, annoying, but much better than not having a telephone.

The major problem was, one never knew, which neighbour was listening
into the call.

Line sharing was not restricted to just houses in the same street. If
one wanted to make an international phone call, one rang the operator
and "booked" the call.

This was because until 1984, Australia only had a total capacity of
eighty [80] simultaneous overseas phone line circuits. [i]

Again, one never knew if the operator stayed on the call after
connecting you to your overseas party. [I'm sure they didn't..]

The upside of course, was that good gossip was spread quickly via the
"party-line" system, therefore, this was undoubtedly the first form of
spoken "broadcast advertising".

Today on the internet, we go about our browsing, emailing, and the three
P's, (peeking, poking and prying), believing that we are safe from
harmful intent; yet hundreds of commercial (NGO) and Government web
sites peer back when we visit their www public personae.

Every-time we search for an item on Google, a little flag is filed away
about our "preferences". "ahh, likes electric golf carts.."

Our media choices are similarly filed away by iTunes and other vendors,
ostensibly to assist us in making choices in the future.

Email from point A to Point O has to pass through numerous routers,
switches and "other" devices to be delivered at the other end.

(Other devices in this instance could be, for example, a computer with
Ethernet in and out ports, looking like a router but designed to collect
all packets and reassemble them. In other words a sniffer that collects
emails.)

The public doesn't realise it but their communications on the Internet
are about as secure as the old "party-lines" from all those decades ago.

Flashback.

1994, Ausnet Services, Sydney Australia with a 64 Kb connection to
Portland Oregon, with 9 GB W3C cache servers located in each of Sydney,
Melbourne, Canberra, Brisbane, Adelaide delivered (comparatively) high
speed Internet to Australians. Provided. they asked for something that
our AI cache engine had calculated they would likely ask for.

Whenever sufficient persons asked for content not in the cache.. Of
course the service slowed down to a crawl.

This is a basic bottleneck function of all networks. For Australians,
the main bottleneck is the "Pond".

Ignoring for the moment, the other ingress/egress cable points in
Australia)

In 1994 Australians shared 256 Kbps [the AARNET service] for their
Internet feeds. Estimates of the time, placed the number of users at
around 180,000 [ii]

If each user was connected at 14,400 kbps dial-up modem speed, (the
majority were, I assure you), then for those users not on Ausnet
services "superior 64 Kb connection, each would have had 1.42 bits per
second.

Fortunately, not all users could get online simultaneously, so the
average user had at least 12 bits per second. Based on an estimated
count of 680 modems servicing Australian Internet in December 1994 and
deducting the usage component of Australia's 36 Universities.

So for early internet adopters who were not at a university, or had
their own ISP, getting online in the mid-nineties was was a frustating -
redial - engaged - redial - engaged - redial - affair.

Of course in 1994, Australia relied heavily on Tasman 2 to New Zealand,
and from thence to Hawaii.

In 1994, Pacrim East and West gave additional bandwidth, but only to
Indonesia and Hawaii again with lots of satellite connectivity to take
up the slack.

Southern Cross was switched on in 2000 and from their web site...

"Like its predecessors, Southern Cross Cable Network has aimed to
provide
the most advanced and innovative service available in our time. We
have learned from the configuration issues of the TASMAN 2 and PacRim
systems - which missed Fiji and only built as far as Hawaii and Guam
- and ensured that we built a network which is fully integrated and
covers all of the key markets in our region."

So as can be seen, even the "experts" get it wrong. Telstra controlled
access by pricing it at astronomical figures. In 1994, Ausnet was quoted
$88,000 per month for a 2 Mb (half circuit) across the pond.

Half Circuit ? Well, that means that the Australian half was connected.
Now the other half circuit had to be purchased form a US carrier as
well.

Today, the average home ADSL connection is anywhere from 256 Kb to 6
Mbits and for cable customers, quite a bit faster (e.g. A user from
Sydney
measured 13762kbps @ Broadband Speedtest.) [although if the head-end is
full of cable subscribers, then most users would be better off on ADSL].

This is where the problem was first discovered.

Any system that requires distribution suffers degradation if all users
turn on the taps at once.

Which is where it gets rather interesting.

Several months ago, I wrote an article that stipulated that multi-media
content (MPEG movies) via cable on an individualized basis was an
impossibility on the current bandwidth capabilities of the carriers.

Since then Telstra has upgraded parts of their network (Melbourne to
Docsis 3.0)

Docsis 1, 2 & 3? Isn't that where the user gets more bandwidth?

Not quite. It's where the cable company get to switch on more users onto
the same cable segment. Although early consumer use of enhanced Docsis
upgraded services may seem to obtain a net speed and throughput
increase.

(The following examples are making allowances for increased Net usage
i.e: non-broadcast content)).

Docsis 1 Less than 3 Mbps (average)
[image]
Docsis 2 less than 5 Mbps
[image]
Docsis 3 less than 5 Mbps
[image]
Each of these solutions in itself is of benefit to the
Telecommunications service provider. The Speed enhancements are
eventually soaked up by the additional users.

More customers are able to be delivered on each platform without
replacing all the cables.

However, for an always on service, where users are utilizing the maximum
bandwidth constantly during peak periods, e.g.: 7:30 pm to 9:30 pm.,
even DOCSIS 3 is insufficient to guarantee service levels to all users
for non-broadcast type content service delivery.

For example, if everyone in Toowoomba, Queensland decided on a Friday
night to download a different HD movie on the BigPond Video on Demand
service, how many of them would be able to stream and watch the movie in
real-time ?

Less than 41%.

In our next article we will examine how Telstra since 2005 have been
working on a way around the bottleneck problem.

References:

i Farewell to ANZCAN Segment E

http://www.southerncrosscables.com/public/News/newsdetail.cfm?StoryID=54

ii Internet Australasia Magazine, Issue 1 Volume 1[Dec 1994], Internet
Providers Guide

Wikipedia Entry on DOCSIS

http://en.wikipedia.org/wiki/DOCSIS

Australian HFC Network

http://www.accc.gov.au/content/item.phtml?itemId=690305&nodeId=e6d85a5da
5875697ff0b57c2ff03a63e&fn=Chap+6+Hybrid+Fibre+Coax+(HFC)+Network.pdf

For History Buffs:

The Magic Wire - THE STORY OF SUBMARINE CABLES

http://p38arover.com/INT/magic.htm

(This is a marvelous site and has multiple links about Telecom/Telstra
cable installation experiences.)

A Short History of Submarine Cables

http://www.iscpc.org/information/History_of_Cables.htm