innsertnamehere
Superstar
I wouldn't be surprised in the least if Metrolinx has to install 10x the presto devices.
|
|
|
rbt
Senior Member
It's somewhat normalized. The faregate installation isn't really related to the card. They could continue using POP without any difficulties. It's an orthogonal decision to attempt to reduce the number of people who opt out of paying a fare and reduce escalating fare enforcement costs.
Either way, TTC has already absorbed the cost of faregates and wiring most stations for networking. To compare the cost with a similar implementation, you would want to add that cost to TTC or remove it from Translink. I opted to remove it from Translink for the above reasons.
Items related to the card itself; readers, backend, control office, call center, ...; is a lower cost.
Project costing is funny. Presto on TTC may include asbestos removal for network upgrades but a several other projects currently on hold will also use those network upgrades (TV monitors, security cameras, etc.)
The group behind Presto seems to insist in ignoring every single lesson learned by Oyster (it's well documented), so as a result we have the 2005 version of Oyster.
Yeah, Presto requires about 3x the equipment for the full implementation (TTC, Ottawa, GO, ...). Economies of scale should make it 2x to 2.5x the cost.
Translink has a surprisingly large number of buses.
Last edited:
rbt
Senior Member
You've underestimated the size of Translink.
Translink has about 1400 buses. TTC has 1800 buses.
Translink has roughly the same number of metro stations as TTC but likely fewer gates
TTC has 200 LRT, Translink as a side collection of ferries, commuter rail, neighbourhood buses, etc.
Both have the similar sized Paratransit fleets (I don't know if either of these will get readers).
GO, Ottawa, and 905 operators are not collectively ~7 times the size of TTC to get that 10x ratio.
Actually, I thought 3x was generous to Presto but Presto will have significantly more backend challenges.
Last edited:
innsertnamehere
Superstar
i was under the impression that the TTC had well over 2000 buses. GO needs tons of Presto machines as well, most of their train stops probably have at 5-10 of them.
rbt
Senior Member
1,857 at the end of 2012.
Vehicle stats are located under the "Easier Access" heading. All buses are accessible, so this is a total count.
http://www.ttc.ca/About_the_TTC/Operating_Statistics/2012.jsp
The artics order isn't going to increase that number by much as some older buses are being decommissioned.
STM seems to have about 1900 buses though obviously no streetcars.
http://www.thetransitwire.com/2012/09/09/montreal-buses-to-get-real-time-information-systems/
Anyway, Translink is very large for a transit organization.
Last edited:
nfitz
Superstar
True, but that doesn't include the 247 streetcars and 200 or so Wheel Trans buses. That's over 2,300 surface vehicles that will need Presto devices ... and who knows how they'll be deploying them in the WheelTrans taxis (and surely they'll have to somehow, or else there will be complaints about discrimination when someone can't use it to pay their fare).
Sure, Translink is big - but not as big as TTC. The point is, that their fare card is being deployed only for Translink. Presto is being deployed for many cities. Quickly looking at the other agencies using Presto, I'm seeing over 3,000 buses (not including WheelTrans-type vehicles). That's over 5,000 vehicles that Presto will be deployed on.
And Skytrain would require less hardware than the Toronto Subway/RT, with only 47 stations compared to 69, and only about 1/3 the ridership.
rbt
Senior Member
Right, and Translink has a couple hundred Neighbourhood buses and Paratransit vehicles too. No denying TTC is larger but translink is not tiny, and may be used as a comparable with adjustments applied.
I never said it was. I gave a ratio of 3x the number of devices for Presto versus Compass Card and have been providing basic vehicle numbers to back up my assertion.
A ratio of 10x is far too much, Presto will not be installed on 15,000 vehicles and 700 stations as currently planned/budgeted.
The ~3x ratio was given for price comparison purposes. Presto should be at most 3x the cost of Compass Card; less for economies of scale.
Last edited:
nfitz
Superstar
Fair enough. 3x is probably reasonable. 4x or 5x might be better.
Still, where does the $100 million number come from? I don't think it's including everything, especially given the $170 million for the fare gates.
On the other hand, some things would be cheaper. The current TTC fareboxes (on surface vehicles) don't have any electricity requirements. I'd think modifying the existing Translink fareboxes for Compass would be a lot cheaper per box than modifying the existing TTC fareboxes.
rbt
Senior Member
GO's bus fleet is shy of 500.
Ottawa has a little under 1000 vehicles.
Mississauga has about 400 vehicles.
Brampton around 300 vehicles.
York Region Transit has around 400 vehicles
Durham Region Transit only seems to have 250 vehicles?
Hamilton has about 200 vehicles
The $170M is cost for both the faregates contract and the Compass Card contracts together.
Compass Card (backend, readers, network, installation, ...) is between $100M and $120M.
Possibly, but faregates on Translink are a change in process unrelated to the card itself. In fact, that decision was made after the fact.
For TTC, I will not be including the salaries of fare enforcement officers for downtown streetcars because it's an orthogonal issue unrelated to the card itself. TTC could have and probably should have done this years ago, or they could have opted not to do it with the new fare card. The card itself didn't cause the change; in fact, the new streetcar design caused the problem.
Wiring the buses for electric fare media is caused by the card, although nearly any change in fare media on TTC would have incurred this cost.
Last edited:
nfitz
Superstar
Surely it's the number of passengers moved per year that creates the required number of readers, and a rough estimate of cost.
According to most recent APTA report, Translink ridership is about 1,090,000 a day. TTC is 2,634,000 a day. Ottawa is 560,000 a day. Go reports about 250,000 a day. Those 3 agencies alone carry 3.2 times more passengers than Translink. And that doesn't include Durham/York/Milton/Hamilton/Burlington/Oakville/Mississauga or Brampton.
I don't think me saying it might be 4x is unreasonable!
Where are the numbers on this from?
I'd expect Translink is going to be a bit cheaper. The card is likely not as sophisticated, not having to handle umpteen different transit systems. They've also gone to a company like Cubic with a lot more experience ... much more sensible that Ontario's bizarre decision to give the contract to Accenture simply because they were the lower bidder for the original tender.
rbt
Senior Member
Not really. A bus can carry 10 passengers per hour or 150, but it still has a single reader. A station needs one read per gate, but a steady flow of 3,000 people per hour is easier to handle than one adjacent to an arena which receives 30,000 people for a single hour a couple times a week.
From this very thread actually, albeit many years ago.
http://urbantoronto.ca/forum/showthread.php/4286-Presto-Fare-Card-(Metrolinx)?p=429991#post429991
http://www.theglobeandmail.com/news...-leaves-big-smoke-in-the-dust/article1659917/
Same article shows the faregates were $90M.
I added $20M to the card project because readers need to be installed into the fare gate machinery. As you know, the fare verification electronics can be swapped out and are separate from the gates but I somehow doubt their price quotes are separated that way.
Last edited:
Electrify
Senior Member
Hamilton only has 200 buses? I'll admit I'm not too familiar with Hamilton's transit system, but I would have assumed they would be second in the GTHA behind Toronto. Never would have guessed it would be smaller than so many suburban municipalities, let alone dead last.
nfitz
Superstar
Nothing to say that TTC is going to implement a single Presto reader on all the vehicles there are going to be 4 readers on the new streetcars.
And then the back-office and transaction time is going to be a function of the total transactions, not vehicles.
It's probably a combination of number of vehicles, stations, and passengers in reality.
It would be interesting to see some realistic estimates of how much it cost, once the system has gone live.
Komiksulo
Active Member
Well, several of the municipalities run their buses on a regional basis... York and Durham, for example.
rbt
Senior Member
That's more related to the vehicle than the number of passengers; but yes, there is some scale with peak passenger load after a fixed cost per vehicle/location. It is peak passenger load though and not overall passenger load. TTC has fairly high mid-day ridership compared to most other North American transit agencies.
It's also typically far cheaper to install a reader into a new vehicle than to retrofit an old vehicle.
10 years ago the backend would have been a huge undertaking. 5 years ago the backend would have been a challenge. Today, not so much unless the software developers really screwed up; which with Accenture is certainly a possibility.
Even open source PostgreSQL (database software) can deal with a 100,000 transactions per second on moderate modern server hardware without breaking a sweat, and that includes committing to a secondary/tertiary locations for high availability purposes. That should be more than sufficient for handling TTC peak load.
Feed it with a RabbitMQ based middleware stack and call it a day.
Free off-the-shelf tools can solve the scaling problem which means they just need to write code that is small and correct and wire the network. Network technology is also a solved problem for this task.
I don't mean to imply that it is a small job; just that it is straight forward with limited technical challenges.
It's like building a 500km long fence. If you can build a 10m fence, then you can build a 500km long fence it'll just take longer and have a few extra bends and perhaps cross a river occasionally so you need to hire the river-fence guy for a few pieces.
Last edited: