10 min read
Passwordless authentication has been the Holy Grail of security for years, but progress has been painfully slow. This does not mean that huge strides have not been made, but unfortunately, most of these developments have been relegated to research labs or professional niches. Until a few years ago, the technology to implement passwordless authentication on a grand scale simply wasn’t available.
However, the industry juggernaut is slowly but surely changing this. There are a few technical, legal and even ethical considerations to take into account, but be as it may, biometric security and passwordless authentication is here to stay.
Biometrics are already changing the game, and they will continue to do so.
Why Go Paswordless In The First Place?
Since this is an engineering blog, I don’t feel the need to explain to a group of security-minded developers the upsides of fast logins. We need not look at the problem from a consumer perspective – all of us are compelled to use a myriad of online services and an ever increasing number of devices. This won’t change anytime soon, and if anything, the number of services and devices we will have to log into will keep increasing.
Of course, there are plenty of ways passwords are being dispensed with, including biometric authentication. From a user perspective, the use of Google, Microsoft and Facebook accounts to log into third-party services works, since the user can avoid password bloat and not have to create accounts for every service and device.
OAuth and OpenID have been used for years to consolidate digital identities, and the standards are employed by some of the biggest names in the tech industry.
Technically, this is not really a passwordless approach, but the average user might not see the distinction.
The pros and cons of using this approach are:
- Easy to implement
- Good security
- Brand name peace of mind
- Dependency on a centralised service
- All eggs in one basket – by compromising one account, an attacker can gain access to others
- Potential security vulnerabilities, beyond your control, can be used against you
- People may be reluctant to use such services due to privacy concerns
Much of this is true of alternative solutions, although it does not apply to security certificates which are usually relegated to business users rather than consumers. The pros outweigh the cons, hence we can already log into countless third-party services using our existing accounts.
How Can Biometrics and Biometric Security Help?
Using biometric authentication systems addresses many issues; there’s no reliance on centralised services, privacy is not a concern, and the user experience is not compromised – provided it’s done right. So, let’s take a look at the pros and cons.
- Fingerprint scanning is quick, cheap and relatively secure
- Voice recognition is easy to use and difficult to manipulate
- Iris scans are very secure and potentially more convenient than fingerprint scanning
- Electrocardiogram technology offers “always on” authentication
- All biometric security methods address privacy concerns while offering good security
- Biometrics are not suitable for all applications
- Cost of deploying biometric security is often prohibitive
- Support is limited to certain platforms and unavailable on most
- Some technologies are still immature
- Biometrics are not a silver bullet – security can still be compromised
Biometrics are not a new concept, or a new technology. Biometric security has been used in many industries for decades, and it’s been a staple of Hollywood script writers even longer. I am sure many readers had a chance to play around with facial recognition and fingerprint scanners on their notebooks years ago – I know I did, and I also know I was not impressed; most of these early solutions were cheap gimmicks.
However, we’ve come a long way since then. More processing power is available, along with vastly superior imaging sensors, and everything is backed by increasingly sophisticated software. This is why some of these technologies are making a comeback, which they’re doing with a vengeance.
Industry Gives Thumbs Up To Fingerprint Scanners
Apple’s Touch ID is probably the most recognisable fingerprint authentication solution on the market, but it’s by no means the only one. Apple opened Touch ID to third-party developers in iOS 8 and proceeded to integrate the technology in new iPhones and iPads, as well as its Apple Pay service.
This is why iOS has a clear lead over Android and other platforms; every new iPhone and iPad will ship with Touch ID until Cupertino comes up with something better.
This does not mean that Android should be written off because an increasing number of Android phones are shipping with fingerprint scanners. The first biometric authentication devices featured small scanners that required the user to swipe their finger over the scanner, but touch-scan units, similar to Apple’s, are becoming increasingly common. It is important to note that this feature is not reserved for expensive, flagship products – even some $200 phones marketed by Chinese vendors feature such scanners.
However, there is still a consideration; Google has not integrated a fingerprint scanner on any of its Nexus devices, although it is rumoured that it originally intended to include it on the Nexus 6 smartphone. In fact, Android Open Source Project (AOSP) provided evidence that fingerprint support was removed from the device. This is not good news for Android developers, as Google usually showcases new technology on Nexus devices and follows up with documentation and APIs, as was the case with NFC support on the Nexus S, or the barometer sensor on the Galaxy Nexus.
Still, this did not prevent vendors from using their own code, with a few types of scanners. But, this is bad news for developers whose hands are tied since there is no standard that would eliminate fragmentation and insure interoperability. Samsung tried to overcome the problem by allowing developers to play around with its Pass API, but this is still not an ideal solution. Motorola tried to do the same four years ago with its old Atrix devices.
A number of hardware manufacturers and developers also released SDKs enabling developers to integrate support for various fingerprint scanners, but the lack of a standardised environment that would reduce or eliminate fragmentation is still a big issue.
It may take a while before we see fingerprint scanners on most phones, but a lot of progress is being made. We went from no scanners on flagship phones to relatively reliable scanners on $200 phones in the space of a couple of years.
But, how useful are they? Are they just gimmicks like first-generation fingerprint scanners on old notebooks?
The technology works, there is no doubt about that, but for the time being applications are limited. Software development has to catch up with hardware, we need more services that can use such solution, and we need more APIs and standards and guidelines from industry leaders (namely, Google). At this point, fingerprint scanners on many Android devices are gimmicks, nothing more.
Overall, fingerprint scanners are convenient, but they’re not an ideal solution. While every fingerprint is unique, there are still some security concerns. Many scanners can be tricked, although it is getting increasingly difficult to pull this off with a simple image. There are alternatives though, including 3D printing, and some morbid ways of doing this, as one security expert pointed out a couple of years ago.
Needless to say, you can’t use fingerprint readers with gloves, an injured thumb, or in other extreme situations. But, these are relatively minor drawbacks.
Microsoft Wants To Look You In The Eye
So, let’s sum up. Android and iOS can already use fingerprint scanners for biometric security, and they are currently underutilised. But what about desktop environments? We can unlock our phones and authenticate payments using biometrics, but we still work on desktops, so how about making them truly passwordless?
Microsoft recently announced Windows Hello and in case you missed it, check out the official Windows blog for a comprehensive overview of this initiative.
This is how Microsoft explains its vision for Windows Hello:
Instead of using a shared or shareable secret like a password, Windows 10 helps to securely authenticate to applications, websites and networks on your behalf—without sending up a password. Thus, there is no shared password stored on their servers for a hacker to potentially compromise.
Windows 10 will ask you to verify that you have possession of your device before it authenticates on your behalf, with a PIN or Windows Hello on devices with biometric sensors. Once authenticated with ‘Passport,’ you will be able to instantly access a growing set of websites and services across a range of industries – favorite commerce sites, email and social networking services, financial institutions, business networks and more.
Windows Hello is a biometric authentication system that will enable users to instantly access their Windows 10 devices, using fingerprint scanning, iris scanning or facial recognition. Microsoft says “plenty” of new Windows 10 devices will support Windows Hello, but, personally, I find one technique particularly interesting.
Iris scanning is one of the methods supported by Microsoft and it has a few benefits over the alternatives. It should be more reliable, and potentially more convenient, than fingerprint scanning. In case you were wondering, this won’t be handled by our webcams or phone cameras – Microsoft wants to use “a combination of special hardware and software” to make sure the system can’t be beaten.
The iris scanner will rely on infrared technology (potentially, near-infrared). This means it will be able to operate in all lighting conditions and see your iris through glasses, even tinted glasses. Hardware designers won’t have to set aside a lot of room on a device to integrate the scanner; it could be integrated right next to the selfie cam on our mobiles, or as an addition to a standalone web cameras used on many office machines today. This means it could be easily retrofitted to existing desktop PCs.
Aside from infrared scanners, Microsoft will also use more traditional biometric security measures such as facial recognition, relying on Intel RealSense camera technology. This should help make Windows Hello more prolific, especially as users upgrade to new notebooks and hybrids based on Intel platforms.
On the mobile front, an iris scan offers several advantages over fingerprint authentication; it can work with gloves, iris injuries are a lot less common than thumb injuries, and it should be much more difficult to beat a consumer grade iris scanner than a fingerprint scanner.
There is another angle to Microsoft’s approach – the software giant won’t store users’ biometric data. The biometric signature will be secured locally on devices and shared with no one but the user. The signature will only be used to unlock the device and Passport, so it won’t be used to authenticate users over the network.
The jury is still out on Microsoft’s biometrics plans and we will have to wait for Windows 10 to see it in action.
What About Always-On Authentication?
While all these technologies might do a good job at replacing traditional passwords, there are emerging concepts that could give engineers more freedom. What if we could dispense with the process entirely, no passwords, no fingerprint scans – nothing?
“Always-on authentication” is the next frontier, and a number of ways of getting there have already been proposed. However, an important distinction needs to be made. Always-on authentication usually refers to machine-to-machine authentication, such as a system of “always-on” SSL authentication, SHH connections, NFC credentials and various networking technologies. These are usually developed to monitor and authenticate financial transactions, thus reducing the risk of online fraud.
There are relatively few solutions for always-on user authentication. One such example is Bionym’s Nymi wristband. It is a wearable device that looks a lot like your average fitness tracker, but it’s more clever than that.
Nymi scans the user’s unique electrocardiogram (ECG). This means that you only need to have the device on your wrist to provide always-on authentication. As long as your heart keeps beating, you’ll be logged in.
If you’re thinking of trying the same trick on the Apple Watch or Android Wear watches, hold your horses, we’re not there yet. The Nymi doesn’t merely track the user’s heart rate like a smartwatch, it actually analyses the shape of the user’s ECG wave, which takes a more sensitive sensor. Smartwatches sound like the ideal hardware platform for this application and, sooner or later, they will be able to do the same thing.
Imagine unlocking your phone, car, office and computer simply by being there and having a pulse? Logging into any account seamlessly, then paying for lunch, doing some shopping on the way home and maybe withdrawing cash from an ATM, all without having to juggle your groceries and credit cards. We’re not there, yet, but we are slowly getting there.
What Does All This Mean For Software Developers And Users?
For the time being, software developers can use off-the-shelf middleware and tokenization to deploy paswordless solutions. One such example is Passwordless, a token-based, open-source framework for Node.js and Express. In case you are interested in how it’s deployed, Mozilla has a comprehensive blog post that explains it.
It will take a while, but biometric building blocks are slowly falling into place. The current crop of passwordless technologies will be augmented, and eventually replaced by biometric authentication.
Many biometric security skeptics including many of my colleagues, don’t believe this will happen anytime soon, but I am an incorrigible optimist; I think passwordless security will be standard by the end of the decade, and this is why: If we merely observe one particular field, be it software or hardware, we will find countless problems with biometrics, many of which I’ve already outlined. However, if we take a few steps back and look at the big picture, if we take a look at new industry trends and the increasing emphasis on personal and corporate security, highly publicised security breaches, privacy concerns – we are bound to see things from a different perspective.
Even so, the elephant in the room isn’t privacy or B2B security, it’s mobile payments.
The volume of mobile transactions in the US is expected to more than double this year to $10bn. By 2018, Bloomberg expects the volume to reach $110bn. On a per-capita basis, the average American consumer will make about $30 in transactions this year, but by 2018 the number will go up to $330 per capita, for every man, woman and child. Assuming the same compound annual growth rate in 2019 and 2020, we could be looking at four digits per capita by 2021.
With that sort of money in play, what do you think?