[ProgSoc] Net on Public Transport

[Marcus Schappi <marcus@xxxxxxxxxxxxxxxxxx>]

Compsoc.s Spectre:
>>Actually, in regards to trains in particular, I came up with this genius
>>solution for the electric variety: Broadband over Power Lines (BPL). The
>>voltage of the overhead power lines used for >>commuter trains is
roughly 25kV, whereas BPL can operate at up to 40kV.

This was my initial idea. (I used it for the basis of my entry for the
2003 UTS Business Planning Competition). It was spurred on because of my
addiction to the net, and the idle time I spent on CityRail Trains.

Once upon a time I used to work for a company called Orbit Australia, and
we did a cost evaluation on just this, the ROI was just not enough to do a
solution like this.

Major Problems with an Overhead Wire Solution encountered were:

-	The lack of Ownership / ability to maintain repeaters situated on
	the overhead wires.
-	The actual cost of the repeaters were more expensive then
	initially expected once the total length of lines was worked out. Eg. In
	some sections of the Newcastle to Central line there is up to 3 tracks.
-	Hardware would still have to be installed on the trains to
	interface with the overhead wires and redistribute as Wi-Fi.

What made an Overhead Wire Solution attractive:

-	The 90% of the CityRail network has overhead wires. (From memory)
	the only line without was the Southern Highlands line?!? (Although our
	resident trainspotter in denial can verify that. Christian..)
-	Low Cost of Bandwidth
-	Speed of Bandwidth

iBurst is attractive because of the ease of installation, all you need is
a low end box with an iBurst Card, a Wi-fi card and a HD with some kind of
OS.

What flavour of linux/unix do those hot-spots in a box run?

How hard would it be (and how long is a length of string) to knit together
a solution that uses debian for an OS, a generic Wi-Fi card, a iBurst Card
(perhaps drivers would have to be reverse engineered), a radius
authentication system that allows external ISPs to hook their customer
base into eg OzEmail, Bigpond, Optus, a content syncing system that
distributed news & entertainment (think video) to the remote boxes, a
modified form of the Squid Caching system that could predict from: a) the
clients that are logged in b) from an overall historical analysis c) time
of day . what sites it should cache; be able to support some form of
geolocation be it GPS or iBurst based, so customers can find out where
their late train or bus currently is (and if it.s worth catching that
other one that doesn.t quite go as directly homeward as your usual bus or
train.

Of course I am now thinking of the privacy implications of the .predictive
caching of webpages based on a users past history.. Everyday we.re slowly
giving up our rights, and I can.t see how this would be any different form
Gmail.s advertisement system reading your email.

The Egg President:
>> How many people on each train/bus/ferry do you expect to use this
>>service at once? I don't think bandwidth will be that much of a
>>problem.

I.m hoping for 20 people per train, but with commoditisation of Wi-Fi
components, you can expect to see them in mobile phones, which creates the
question of VOIP. What sort of drain is VOIP on a 1.5Mbit system?

>>You'd probably be bettor off streaming your digital TV from one central
>>point as opposed to trying to incorporate it into all your little
blackboxes.

Bandwidth.

>>And if this is going on public transport is there any reason it needs
>>authentication? Why not make it free?

Just how much do you like advertising?
It goes back to the age old question that seemed to be missed during .the
bubble. WHO PAYS and WHY.

>> perhaps you could have people go through a portal that links to news
>>sites and such, and that way you'd know what you'd need to be leading
>>people to content that's already cached.

Kind of reminds me of my crappy .3. service. As Ck would say, the beauty
of the internet is that people are able to choose what content they want
to view.

>>And what is the new hack at Queensland Uni you are talking about?

From memory the hack amounted to hacking the firmware of a Wi-fi card and
broadcasting on all channels simultaneously to a hotspot.

Vik:
>>So you would be charging the commuters for use of said system?

Hell yes! Do you think that CityRail is going to pay for it? If you are
one of those poor people that are forced to take a certain undesirable
form of public transport to uni, work or the like you will still do so
even if there is no internet on the form of transport.

>> set-and-forget servers which have to be mostly bulletproof.

Yes that brings me to an interesting question of how would one remotely
administer all these boxes? Updating them would be tantamount to upgrading
the firmware of a Mars Rover.

>>a repeating news-bulletin type thing to play as a backup when the signal
>>drops out. Copyright stuff about re-broadcasting TV signals should be
>>considered too.

Lets step into a hypothetical world where you are Kerry Packer. You own a
free to air network, would you not want your Television and Ads
re-broadcast to those scores of people who travel by public transport each
day during prime time? (Where you charge the most for ads and pay the most
for content?) All these people on public transport carrying symbian java
phones with bluetooth, and PDAs / laptops with Wi-Fi, all ripe for the
streaming of .mp4 video.

What happens then when you know what.s on the box, have internet access at
the same time? So say Current Affair does a report on a new type of Maths
Computer Tutor. you can now link to the site directly from the show being
streamed. (Profit.)

~ Marcus



-
You are subscribed to the progsoc mailing list. To unsubscribe, send a
message containing "unsubscribe" to progsoc-request@progsoc.uts.edu.au.
If you are having trouble, ask owner-progsoc@progsoc.uts.edu.au for help.