Ahh springtime, flowers and the annual Apple Platform Security (APS) update. This year’s version has many Apple Pay housekeeping changes. Previous versions put everything Apple Pay in a single section. In keeping with Apple spinning out iOS 15 Wallet app as a separate identity, Wallet has its own separate section now, covering all the things Jennifer Bailey unveiled at WWDC21: hotel-home-office keys and ID in Wallet. The Apple Pay section adds a new category for Tap to Pay on iPhone with some interesting bits.
The Tap to Pay on iPhone servers manage the setup and provisioning of the payment kernels in the device. The servers also monitor the security of the Tap to Pay on iPhone devices in a manner compatible with to the Contactless Payments on COTS (CPoC) standard from the Payment Card Industry Security Standards Council (PCI SSC) and are PCI DSS compliant.
The Tap to Pay on iPhone server emits decryption keys to the Payment Service Provider after validation of the integrity and authenticity of the data, and after verifying that the card read was within 60 seconds of the card read on the device.
What’s interesting to me is that Tap to Pay on iPhone servers are providing a seamless payment reader experience in the same way that Apple Pay servers provide a seamless pay experience. It just works, from setup to use, the same tight integration allows payment service providers to focus on POS app development and forget about the hardware because Apple Pay takes care of everything. As Junya Suzuki tweeted recently, a lot of payment reader hardware is suddenly junk compared to what iPhone is providing with tight mobile integration and Tap to Pay servers on the backend. Now with Tap to Pay apps on the horizon, good thing that iOS 15 Wallet expanded the secure element max to 16 ain’t it?
Speaking of Wallet, this separate section covers all things “access credential” related (hotel-corporate-home-car-student ID) with App Clips suggested for provisioning multifamily home keys. Transit now includes eMoney cards (or is it e-Money, Apple seems confused about it just like Express Mode vs Express Transit) and IDs in Wallet is covered in detail. There is also an intriguing iOS 15.4 Wallet security tweak:
In iOS 15.4 or later, when a user double-clicks the side button on an iPhone with Face ID or double-clicks the Home button on an iPhone with Touch ID, their passes and access key details aren’t displayed until they authenticate to the device. Either Face ID, Touch ID, or passcode authentication is required before pass specific information including hotel booking details are displayed in Apple Wallet.
It sounds almost exactly what we already do with regular Apple Pay cards. Perhaps keys and passes only show a generic icon and checkmark with Express Mode with the double-click + authentication required for show details…it’s not very clear.
The whole security expert thing reminds me of what my uncle the doctor (who ran a medical research lab at Columbia University) used to say about his disdain for pharmaceutical companies, “They don’t want to cure you, they just want to keep ‘treating’ you with their medicines.” Human nature never changes. The gist is that EMV Express Transit Mode will always be a thorn in Apple Pay’s side because the security is up to the card companies.
The document is worth your time is you have any interest in Apple Pay and Wallet.
(The) Digital Markets Act will…require companies designated as gatekeepers to ensure effective interoperability with hardware and software features they use themselves in their ecosystems. This includes access to NFC for mobile payments.
Today’s case addresses a conduct by Apple that has been ongoing since Apple Pay was first rolled out in 2015 <sic, 2014 actually>. This conduct may have distorted competition on the mobile wallets market in Europe. It prevented emergence of new and innovative competition that could have challenged Apple.
Both pieces miss important context surrounding the debate however…and with this issue context is all, especially how Apple Pay is playing out in other global markets. Most of what follows I’ve covered in earlier posts but hope to pull the various issues together in one post. Yet again, we kickoff with an updated Apple Pay diagram.
The so called Apple ‘NFC chip’ is not a chip at all but a hardware/software sandwich. The Apple Pay ecosystem described in iOS Security is a collection of tightly integrated polished pieces: Secure Element, Secure Enclave, NFC Controller, Wallet and Apple Pay Servers, all wrapped into a slick, easy to use UI with a final security wall of ‘secure intent’, a double-click side button hot-wired to the Secure Element. This approach has been so successful that people divide mobile payments history into pre-Apple Pay and post-Apple Pay eras.
Apple Pay has a very simple rule: any card that loads a Java Card applet into their embedded secure element (eSE) has to reside in Wallet app. The maximum number depends on how many Java Card applets it can hold at any one time, the previous limit was 12, the iOS 15 Wallet limit is 16 cards. Developers have two ways to access iPhone NFC: 1) Core NFC framework for NFC operations that don’t use the secure element, 2) Secure Element pass certificates for NFC operations that need secure element transactions (payments, keys, ID, passes). Any developer who wants to run applets in the eSE has to apply for a PassKit NFC/Secure Element Pass Certificate. This is covered by NDA but a company called PassKit (not Apple) gives us an idea what Apple’s Secure Element Pass guidelines are:
Apple care a great deal about the user experience. Before granting NFC certificate access they will ensure that you have the necessary hardware, software and capabilities to develop or deploy an ecosystem that is going to deliver an experience consistent with their guidelines.
The end to end user experience, the whole reason behind the success of Apple Pay. But this gatekeeping is what riles banks and financial service providers who want to load their applets into the secure element without the Apple Pay gatekeeping, without the Apple Pay ecosystem and without the Apple Pay commission. They want to do their own transactions with their own app for free. This is what the EU Commission means when Vestager says: “Evidence on our file indicates that some developers did not go ahead with their plans as they were not able to to (sic) reach iPhone users.” It should read: when they were not able to reach iPhone users for free. Either the developer didn’t apply for a Secure Element Pass, didn’t pass the certification process, balked at Apple’s certification conditions, or couldn’t agree on Apple Pay commission rates.
Secure element gatekeeping is not new, it is an essential part of the secure element system:
A Secure Element (SE) is a microprocessor chip which can store sensitive data and run secure apps such as payment. It acts as a vault, protecting what’s inside the SE (applications and data) from malware attacks that are typical in the host (i.e. the device operating system). Secure Elements handle all sorts of applications that are vital to our modern digital lives…
Mobile Payments Here, the Secure Element securely stores card/cardholder data and manages the reading of encrypted data. During a payment transaction it acts like a contactless payment card using industry standard technology to help authorize a transaction. The Secure Element could either be embedded in the phone or embedded in your SIM card.
Lifecycle management It’s crucial that SE-embedded devices are secure throughout their lifecycle. That’s why Secure Elements need to have an end-to-end security strategy. It’s no use developing a robust security solution for a device which becomes obsolete after a period of use. This is why Secured Elements can be updated continuously to counter new threats.
Few people, especially a PayPal or EU Commission vice president, discuss the crucial secure element lifecycle management aspect. It’s not convenient for them to say the secure element ‘gatekeeper’ is responsible for keeping it secure. Far more convenient for their arguments to omit this, portray gatekeeping as unnecessary and gatekeepers as evil. In the end however, Apple has to maintain secure element updates from the various licensed secure element providers (EMV,FeliCa Networks, MIFARE, and so on) if secure payments are going to work at all This is what people who say, ‘it’s my device, we should be able to use NFC how we want,’ do not understand.
People also forget that nothing is free, you get what you pay for. With Apple Pay as gatekeeper, users get simplicity, innovation and feature updates. Simplicity: users get NFC they can use out of the box without Android-like NFC complexity such as secure element positions and obscure express mode settings.
Innovation: Apple Pay has features like Global NFC. iPhone and Apple Watch are the only smart devices that come with FeliCa built in as standard to use in Hong Kong or Japan, while Android limits functionality by market region. It’s astounding that Android, not even Google Pixel Android, has matched this basic functionality yet. We’re seeing more innovation as Ultra Wide Band (UWB) extends Wallet functionality to include ‘Touchless’ car keys and eventually, UWB enhanced automatic card selection as you approach the reader; more helpful than you might think.
Japan is key to understanding what’s really going on in the Apple Pay monopoly debate. Japan was the first market with an established mobile payment platform in place, long before mobile EMV contactless payments took off in Europe. iPhone also has a much larger marketshare in Japan than it does in Europe. It’s a shame people pass up the opportunity to learn from the successes and failures here.
So what’s the EU Committee vision for ‘open NFC’? I think it’s a rehash of the secure element wars when carriers locked mobile payment services to SIM contracts. In 2013 Google incorporated SimplyTapp HCE (Host Card Emulation ‘secure element in the cloud’) technology as a NFC ‘workaround’ to ‘free’ NFC from the evil clutches of mobile carriers. Sound familiar? Android NFC has never been right since.
How little things change, swap ‘evil mobile carriers’ for ‘evil Apple’ and you have the same self serving ‘open’ vs ‘closed’ NFC chip nonsense that people are debating today. FeliCa Dude, the ultimate industry insider who has experienced it all, said it best: ‘It’s all eSE or nothing now.’
And yet we now have Île-de-France Mobilités (IDFM) turning back the clock, circumventing the eSE on NFC equipped Android devices and going all in with HCE for IDFM’s Smart Navigo service for Android. To me this says all you need to know what European priorities are regarding the ‘open NFC’ model: eliminate eSE gatekeepers by forcing the less secure network dependent HCE as a required option. Good luck with that. From a transit perspective, based on Mobile Suica user experiences, I don’t think HCE Smart Navigo will be a smooth ride.
The EU Committee ‘open NFC’ vision might look ideal…to Apple Pay competitors. Regular users however, will have to deal with the ugly reality of multiple NFC apps, multiple NFC secure element modes and clashing updates that cancel out NFC services. Apple Silicon eSE space is limited to 16 cards. If that sounds like a lot now, wait until you have credit cards, transit cards, home, car and office keys and ID installed along with ‘open’ NFC apps wanting their own eSE space too. Services will be squeezed out forcing the user to intervene. If the EU Committee thinks this environment fosters competition and innovation while growing mobile payment use, dream on.
Japanese tech journalist Junya Suzuki has covered NFC mobile payment developments in Europe, America and Japan for over 2 decades. He doesn’t think the EU is playing an even hand here, in his opinion Samsung and Huawei would never face the scrutiny that Apple now faces. In typical European cultural fashion, EU motives pay lip service to fair open markets while playing an underhanded game of chess to make Apple do what EU banking interests want Apple to do. In other words, a double standard.
What does Apple need to do? I’ve always said that Apple needs to make the Secure Element Pass application process as transparent as possible. Keeping the blackbox NDA process as it is now makes Apple Pay a target, increasingly difficult to defend the status quo. Secure Element access on the level of Core NFC is a long shot, the very definition of a secure element means there has to be a developer certification process similar to EMVCo, FeliCa Networks, MIFARE, Calypso Networks Association, etc., that protects the privacy and business interests of all parties. But it would be great if there is a middle way where Apple can securely open things up for iPhone as a digital wallet, and iPhone as a payment terminal. We’ll see if Apple has anything to say about the subject at WWDC22.
Recommended reading: Ruimin Yang’s wonderfully detailed analysis, “Apple Pay monopoly, are we really comparing ‘Apples’ with ‘Apples?“outlines the entire Apple Pay system architecture, how it compares to other digital wallet platforms, (Google Pay, Samsung Pay) and what ‘open vs closed’ means in the ‘Apple Pay is a monopoly’ debate.
The April 19 launch of SBI Neobank Mastercard debit card support for Apple Pay was a bit unique: the first time that a plastic issue Japanese debit card came to Apple Pay and the first Apple Pay Japan debit card supporting the FeliCa iD payment network. Another interesting aspect is that only the Mastercard version supports Apple Pay, the VISA version is plastic only with VISA Touch (EMV contactless) support.
There are plenty of bank app issue digital only debit cards from JCB, Mizuho, MUFG and others on Apple Pay. These all work on JCB’s QUICPay (FeliCa) and J/Speedy(EMV) payment networks. Apple Pay Japan supports many different mobile payment network cards thanks to Mobile FeliCa support, by far the largest selection of Apple Pay payment networks in the world: EMV (VISA, Mastercard, AMEX, JCB), iD, QUICPay, Suica, PASMO, nanaco, WAON. But VISA issue debit cards are not supported even though there are many, not a single one on Apple Pay.
Wasn’t this taken care of by the May 2021 Apple and VISA JP agreement? For credit cards yes, one year later they are still at odds over FeliCa support in debit cards. VISA Japan brand debit cards are VISA Touch EMV contactless exclusive, single mode cards. VISA JP credit cards are dual mode EMV/FeliCa for plastic and smartphones, but not debit cards. We don’t know the reason but debit cards deifintely fit the budget customer category while credit cards come with credit checks, perks and card membership fees for upscale cards.
As an easily available budget card, VISA cuts costs by dumping the dual mode EMV/FeliCa IC chip and transaction fees for the convenience of using FeliCa iD/QUICPay payment networks. In other words VISA keeps all transaction fees for themselves while marketing the shit out of VISA Touch as the greatest thing since…whenever.
All of the other card brands in Japan have dual mode NFC as standard. Not VISA, they’re playing the long game of eliminating FeliCa payment network competition. This stupid polarizing single flavor NFC position only served to give QR Code payment networks (PayPay, Line Pay, etc.) a huge opportunity that they smartly played. End result: more payment network competition than ever before.
Apple on the other hand has a very simple rule for all Apple Pay Japanese issue cards: they must support FeliCa and all EMV cards are global NFC dual mode. Was this the price for adding FeliCa support to Apple Pay? Perhaps, I think it’s more to do with the Apple Pay vision of removing complex and confusing hardware choices, the Google Pay Japan mess, for standard ‘just works everywhere’ NFC. Has this been successful? Very...just ask Suica.
Reece Martin posted an interesting video, So you built the wrong transit system, that examines the American penchant for building cheap light rail systems that don’t make long term sense. Public transit is a waste of money to Americans with money, so cheap is only way to fund and build public transit infrastructure. The problem is this cheap short term thinking costs more money in the long run. It’s a ‘one size fits all’ mentality.
But as Reece points out, systems can evolve from humble beginnings. Many private Japanese rail lines started out as street trams (that evolved from horse trams) but evolved into the heavy duty regional rail lines we have today. Fare system have evolved too, from paper, to mag strip, to IC smartcard and now mobile devices.
Transit fare systems in America suffer from the same short term cheap thinking, on full display on the MTA OMNY system, the world’s first EMV only open + closed loop fare system. When it’s completed in 2023, barring more delays, MTA will have farmed out every aspect of their fare collection and OMNY transit card issue to banks.
Not to rehash points I already made about OMNY, but Reece’s wrong transit system analogy struck a chord. And unlike rail system evolution, once the transit fare system in locked into the bank payment card infrastructure, from technology (EMV) to payment network processing (VISA, mastercard, AMEX, etc.), it will be extremely difficult, if not impossible to change anything later on.
But why is America so short sighted when it comes to public transit, never investing in a long term self-sustaining viable business model? I ran across an interesting take that explains it neatly. The USA will never have a transit platform business because public transit is a welfare and jobs program, not a self-sustaining business model:
Public transportation in the US is generally very bad and very heavily subsidized. It’s cheap because extremely little service is being run, and the government picks up most of the bill.
Public transportation in the US is less of a way normal people get around, and more of a welfare program and jobs program. Even in places where public transportation is a way normal people get around, e.g., NYC, it is run more like a jobs program than an essential public service.
Open loop fare systems are also vulnerable in new ways nobody predicted: imagine the mess if payment networks go down in a cyberwar, à la the Moscow metro when digital wallets and bank payment card networks were suddenly and omniously turned off. In the case of OMNY where, unlike Moscow metro, everything is EMV payment networked…there is no backup in-house payment settlement system, there is no plan b.
In other words not only is OMNY EMV one size fits all. it’s all or nothing.
After a test phase, in 2022, iPhones and Apple Watches will be able to replace the plastic pass distributed by IDFM (in 2023). “We cannot yet give a precise date, because it depends on the progress of Apple’s developments in Cupertino. But this time, for sure, it will be done, “says Laurent Probst, CEO of Île-de-France Mobilités. The contract is due to be voted on this Thursday at IDFM’s board of directors…
The contract between IDFM and Apple is spread over a period of five years, with a total budget of up to €5 million dedicated to the development of new services. A budget equivalent to that allocated to Android service developments operated by Samsung with IDFM.
The contract with Apple is due to be approved by IDFM directors the week of February 20, we can thank the 2024 Paris Summer Olympics for breaking the Smart Navigo on Apple Pay logjam. Le Parisien has regularly criticized IDFM’s slow rollout of mobile services: “The modernization of the ticketing system in force on public transport networks in Île-de-France is not a long quiet river.” A timeline is helpful to understand the stalemate.
October 2017: Smart Navigo mobile was announced for 2019 launch. At the time IDFM said, “Unfortunately, it won’t be possible for iPhone owners to use the service since Apple does not yet allow third parties to access the NFC secure element in their phones. However, we are happy to explore the possibilities with Apple to offer the same service to all Paris public transport users.” In other words, IDFM wants to bypass Apple Pay Wallet and do everything in their own app.
September 2019: Smart Navigo launches on smartphones using an Orange SIM card, and on Samsung devices.
February 2022: Le Parisien reports Smart Navigo on Apple Pay will launch in 2023, IDFM confirms on Twitter and also announces EMV open loop support coming in 2024 in time for the 2024 Paris Summer Olympics.
French journalist Nicolas Lellouche independently confirmed the Apple Pay Navigo 2023 launch directly with IDFM and posted some details. Expect direct adding in Wallet app with Apple Pay recharge, similar to Suica, PASMO, Clipper, TAP and SmarTrip. An updated ViaNavigo app will provide extra features for commuter passes and more service options.
French reaction on Twitter was interesting and varied. People complained about the long lag getting Smart Navigo on iPhone but the equally long delay getting Smart Navigo on all Android devices, not just Samsung Galaxy, is more interesting and revealing. IDFM has spent a lot of time and expense working with Calypso Networks Association, the transaction tech used for Navigo, to develop the less secure network dependent Calypso HCE ‘cloud’ secure element approach. It flies in the face of where payment transaction technology has been going with eSE as standard hardware on all modern NFC devices. It’s almost like Ferdinand de Lesseps digging a sea level Panama Canal when a lock-and-lake canal was the better technical choice all along.
As for Android Calypso HCE performance vs Apple Pay Navigo Calypso eSE performance, I suspect the network dependent HCE on Android will be problematic. It will certainly be problematic, and challenging, for non-Apple smart wearables. If there is anything the bad user reviews of Suica App tell us, it is that network connections in station areas and on trains are never reliable and Android NFC adds layer upon layer of support complexity. No network = no HCE service, it’s that simple. Apple Pay Navigo will work without a network connection, just like all transit cards on Apple Pay, and will work great on Apple Watch too.
For this reason IDFM has to focus all of their system resources on the much more complex Android launch this year. They could certainly launch Apple Pay Navigo sooner if they really wanted to, but it’s better to do these things one platform at a time.