As co-chair of the W3C community group aiming to offer to web developers the possibility to access to services provided by hardware token, I am receiving some questions on a regular basis about where does this work go …
Well. Executive summary. The good reasons for allowing a web app to access to secure services stored in a harware token, and the possible ways to implement that in browsers are ready. But this is still not in the W3C planet. This is in a form of a report, edited by Sébastien Bahloul, a Morpho guy, and discussed with W3C Community Group members.
In details. The good reasons for allowing a web developers to access to keys stored in a hardware toke, or to trigger a signature which can not be repudiated are detailed in the report. There are some specific industry examples, such as government e-services, or e-banking services, or commercial transaction, which requires legal binding, such as online signature. The potential users of this feature are legions. Basically, the european regulation, named eIDAS “regulates electronic signatures, electronic transactions, involved bodies and their embedding processes to provide a safe way for users to conduct business online like electronic funds transfer or transactions with public services”. To deploy such services on the web, the web developer needs to have some mean to access hardware token (or the web will miss that digital european trust promise). Other countries such as Bolivia, Uruguay, Argentina and Peru are also requiring similar technology.
The technical aspects. The technical proposal embedded in this report is made of two technical features. First. A way to implement the W3C Web Crypto API in hardware token. this is to allow the generation and the usage of a cryptographic key inside a token belonging to the user. Second. A way to digitally sign a transaction with a a key, again stored in a hardware token, and performing the signature confirmation via an interface the user can trust. Those two services are some of the building blocks to have a trusted web, where the user is in control of the credentials used to cipher or sign some data.
So what is wrong ? Well. This set of usages and technical feature were presented to a large group of W3C members during last W3C TPAC. And, nothing amazing happened. The browser makers were kindly requested to have a look at it. But they demonstrated low interest, while this topic has been discussed since september 2014. There might have a cultural problem here behing the slow progress of this topic in W3C. The online access to european government services is not a priority for the major browser makers. In addition most of the actors of the security have managed some hacks to be able to use smart cards or hardware token, like plugins. But this era is over, as plugins maintenance and attacks are getting more sensitive.
And what is next ? Next is about gathering the companies and countries interested in that feature, and start to demonstrate W3C that there is an important question here : do we want the web to get in the secure services, as requested by online signature and government services ? So if you are part of the actors believing this web feature is key, join the Hardware Based Secure Services CG, so that we can collectively work on creating a Working Group in W3C…
Thanks a lot for your explanations, Virginie.
I tried to relate this situation, applied to the Bolivian context, in the following article (Spanish):