Hidden Assumptions

Jonathan Seybold said it best in his Computer History Museum interview video, many arguments can be easily demolished by pulling out the hidden assumptions. In our attention span challenged social media era it’s all too easy to believe things at face value. Few people invest time and brain energy to analyze and question arguments to find and examine hidden assumptions.

A reader of this blog might come away thinking I am not a fan of open loop transit fare payments and despise EMV contactless and QR Code payment technology. That would be a mistake. I don’t hate them, everything has its place. I simply don’t agree with ubiquitous assumptions that EMV or QR or open loop are cure alls for every transit fare payment situation that they are praised to be…usually because ‘everybody uses’ bank issued contactless payment cards or smartphone payment QR apps. It’s a one size fits all mentality that blinds people from seeing hidden assumptions. It’s very important to see how all the pieces, seen and unseen, fit together. After all, transit companies and their users have to live with transit infrastructure choices for decades.

In a recent twitter thread Reece Martin thought it would be nice if Canada had a nationwide transit card. This is something Japan has had since 2013 when the Transit IC interoperability scheme was put in place that made the major transit IC cards compatible with each other, but they did this without changing the hardware. The various card architectures were left untouched and linked with system updates, a use-the-same-card backend solution. China on the other hand created a national transit card with the China T-Union • PBOC 2.0 standard that replaced all older transit cards with locally branded T-Union cards, a get-a-new-card hardware solution.

A nationwide Canadian transit card is a great idea but as Samual Muransky answered in the same thread, why bother with ‘obsolete’ dedicated transit cards when everybody uses EMV contactless bank cards and EMV is the new standard. Let’s examine some hidden assumptions at play here.

Assumption #1: Everybody has contactless credit/debit cards
The open assumption here that everybody has bank issued credit or debit payment cards is not the case and varies by country, demographics, age, etc. Most people in some countries do, but even so there will always be people who don’t. Transit cards always have the advantage of being available at station kiosks to anyone with cash.

Assumption #2: because of assumption #1 open loop (credit/debit cards) is better than closed loop (dedicated ticketing) for paying transit fare
The hidden assumption is that open loop covers everything but it does not. Specific transit services such as individual commuter passes, discounted fares for disabled/elderly/children are practically impossible to attach and use with bank payment cards. The best that transit systems and payment networks can do with open loop is fare capping or special discounts when applied universally. The age-old pay ‘x’ times and get one free concept. Open loop works best for occasional transit users.

The limitations of open loop on large complex transit systems like Transport for London is easy to see. Despite a long campaign to eliminate the venerable Oyster transit card and migrate users to EMV open loop, TfL threw in the towel and upgraded the Oyster system recently. To date TfL has not offered a digital version of the closed loop Oyster card. In short, dedicated transit cards will always be with us.

Assumption #3: EMV contactless is the NFC standard
The NFC Forum recognized long ago that credit card companies and transit companies have different needs and objectives. To that end the NCF Forum has 2 basic NFC standards, one for contactless payments (NFC A) and one for transit (NFC A-B-F). All NFC devices must support NFC A-B-F for NFC Forum certification.

Assumption #4: EMV contactless for transit is safe and secure
There are many hidden assumptions packed into the words ‘safe and secure’: not everybody agrees on what safe is and what level of security is secure. Things also change depending on the situation and the design. I have covered transit gate reader design in many other posts but recap some basics here.

Steve Jobs famously said that designing a product is a package of choices. I have often said that EMV contactless is supermarket checkout payment technology but that’s not a put down, it’s the truth of what EMVCo were aiming for when they grafted NFC-A to their EMV chip for contactless cards.

Because of wide deployment with no direct control, the original EMV contactless spec had a latency window to work reliably even with crappy network installations, and the slow speed has sometimes been cited as a security risk. NFC-A (MIFARE and EMV) transaction speeds are rated for a theoretical 250ms but are usually 500ms on open loop transit gates. Suica is always 200ms, often faster. The speed gap is due to gate reader design, the network lag of centralized processing vs local stored value processing, and the different RF communication distances for NFC-A and NFC-F. JR East presentation slides explain the transaction speed differences.

  • Japanese station gates are designed to be capable of 60 passengers per minute. To do this the conditions are:
    • Processing time of fare transaction has to be within 200ms
    • RF communication distance is 85mm for physical cards and smartphones
  • European station gates are designed to be capable of 30 passengers per minute:
    • The processing time takes 500ms
    • RF communication distance is 20mm for physical cards, 40mm for smartphones
016l
Presentation slide from the NFC Forum Japan meeting, July 2016
018l
Presentation slide from the NFC Forum Japan meeting, July 2016

The Suica transaction starts from the 85mm mark while MIFARE and EMV contactless cards start at the 20mm mark. Because of the greater RF communication distance Suica transactions start much earlier as the card travels toward the reader tap area. It you look closely at the 2nd slide you can see that smartphones have a slightly earlier EMV/MIFARE RF transaction starting at the 40mm mark (the 1.1A/m boundary) due to the larger smartphone antenna, physical EMV cards with smaller antennas are limited to 20mm. This is why smartphones seem faster than physical cards on NFC-A gates. Suica physical cards have a larger antenna and the same RF transaction distance as smartphones.

NFC-A transaction speed is slower because it has to be on top of the reader before it can start. This is also the limitation with optical based QR and bar codes, the transaction only starts when the smartphone screen is close enough to the reader for an error free scan. Transit gates using these technologies are not designed for smooth walk through flow.

The speed difference is clearly seen on the Nankai VISA Touch open loop gates: the transaction starts when the card is physically on top of the reader:

Here is Suica style transit gate for comparison:

One of the smart things Nankai is doing in the test phase (limited to a few key stations) is keeping EMV/QR gates separate from standard FeliCa gates. This is practical. Regular users go through the faster regular gates, the occasional open loop or QR users go through slower EMV/QR gates. Keeping different readers separate and clearly marked helps keep walk flow smooth and crowding down at busier stations. The Nankai program has been put on pause for another year due to the collapse of inbound travelers in the COVID pandemic. It’s a trial run as Osaka area transit gear up for an anticipated inbound travel boom in connection with Expo 2025, that may, or may not pan out.

The Nankai VISA Touch gates are designed for physical cards, Apple Pay works but without Express Transit. That’s a plus as Apple Pay EMV Express Transit on TfL and other open loop systems (OMNY) has come under scrutiny for a potential security risk with VISA cards that allows ‘scammers’ (in lab settings) to make non-transit charges to Apple Pay VISA cards via Express Mode, something that is not supposed to be possible.

Timur Yunusov, a senior security expert at Positive Technologies…said a lack of offline data authentication allows this exploit, even though there are EMVCo specifications covering these transactions.

“The only problem is that now big companies like MasterCard, Visa and AMEX don’t need to follow these standards when we talk about NFC payments – these companies diverged in the early 2010s, and everyone is now doing what they want here,” he said.

Security researcher: Flaw in Apple Pay, Samsung Pay and Google Pay makes fraud easy for thieves, Techepublic

In other words, Apple removing Apple Pay bio-authentication to promote EMV Express Mode for open loop transit puts Apple Pay at the mercy of lax card network payment operation practices who don’t follow their own rules. Not that it’s a real problem in the field but accidents do happen, such as this incident on Vancouver BC TransLink that a reader forwarded:

Just a moment ago, I nearly got dinged on my CC while sitting on a high seat near a door which is where one of the validators are located. The validator picked it up from the backside rather than the front side where the tap area is located. Also, somehow, my iPhone authorized the transaction when I only want to return to the home screen instead.

If the open-loop was implemented in a way where the card must be pre authorized before the card can be tapped at a validator, it wouldn’t get me in a situation where I need to deal with customer service to dispute some charges. Good thing this time, transaction was declined so nothing related to this charge showed up in my account.

Smartphone users be careful around the backside of Vancouver BC TransLink pole readers

Open loop is only part of a larger picture
Canadian transit would certainly benefit from a Japanese transit IC system approach with compatibility on the backend, or even the China T-Union approach of a national card spec that is locally branded but works everywhere.

To come back to the beginning, my point isn’t about slamming EMV or QR open loop transit, just the assumptions that they solve everything. They have their place in intelligently designed fare systems but only constitute part of the larger transit fare system picture. And as I have pointed out many times, card companies have little interest in improving the EMV standard for transit needs. They want to capture transit fare business without investing. The focus will always be the supermarket checkout lane that EMV was designed for.

There will always be a risk involved when ignoring the hidden assumptions of EMV open loop as a one size fits all solution. Dedicated transit cards will always be necessary. Every transit system is unique and deserves the best solution for the transit company and the users they serve.