Pat Farrell wrote: > Nearly ten years ago, when I was at Cybercash, we worked with Mondex and > other smartcard vendors who also said "as soon as we have infrastructure" > > Something tells me that soon is not gonna happen in what I would > call soon. Smartcards (the smart part) were moderately interesting > when there was no networking. We've been at ubiquitous networking > for many years. > > While he was at Cybercash, Ellison was awarded US Patent 6,073,237 > "Tamper resistant method and apparatus" which is precisely > a network based, software only smartcard.
my characterizations of smartcards from the 80s ... was that they were targeted at the portable computing market segment. however, the technology was only sufficient for the chip ... and there wasn't corresponding portable technology for input and output. as a result you saw things like the work in ISO for standardizing interface to the chip ... so the chipcard could be carried around and interop with fixed input/output stations. in the early 80s, you saw the advent of PDAs and cellphones with portable input/output technology that sort of took over that market. which would you prefer a portable computing device with lots of application and data where you had to go find a fixed input/output station to utilize the device .... or a similar portable computing device where the input/output was integrated? in the 90s, anne & I were asked to spec, design, & cost the infrastructure for a mondex roll-out in the US ... aka it wasn't the mondex card per-se ... it was all the rest of the infrastructure and dataprocessing required to support a mondex infrastructure (from the mondex international superbrick on down to loading/unloading value on the chip). one of the financial issues with mondex was that most of the float & value was at mondex international with the superbrick; in fact later on you saw mondex international making inducements to various countries where they offered to split the float. this was about the time several of the EU central banks made the statement that the current genre of stored-value smartcards would be given a couple year grace period allowing them to establish an infrastructure ... but after that they would be required to pay interest on unspent value in the card (would have pretty much eliminated the float value at higher levels in the operational stream). that was coupled with the fact that it had a fundamental offline design point ... i.e. the value was held in the chip ... and could be moved between chips w/o having to go online ... becomes something of an anachronism if you have ubiquitous online access (as you've observed). mondex also sponsored a ietf working group looking at possibly application of mondex transactions in the internet environment. that really represented a difficult undertaking being a shared-secret based infrastructure. the working group somewhat morphed and eventually turned out ECML and some other stuff ... some recent RFCs .. XML Voucher: Generic Voucher Language http://www.garlic.com/~lynn/rfcidx13.htm#4153 Voucher Trading System Application Programming Interface (VTS-API) http://www.garlic.com/~lynn/rfcidx13.htm#4154 which evolved out of the work on ECML (electronic commerce markup language), which in turned started out with working group somewhat looking at adapting Mondex to Internet transactions. Electronic Commerce Modeling Language (ECML) Version 2 Specification http://www.garlic.com/~lynn/rfcidx13.htm#4112 some of that chipcard technology can be applied to an electronic "something you have" authentication technology ... where it is difficult to compromise and/or counterfeit a valid chip. this raises something of a perception issue ... if you stick with the protable computing device model ... then the chipcard has a bunch of capability that is redundant and/or superfluous for somebody with a cellphone/pda. If you go with purely the (hard to compromise and counterfeit) "something you have" authentication model in an online world ... then KISS (or Occam's Razor) would imply that most of the features associated with the earlier smartcard model are redundant and superfluous (and might actually pose unnecessary complexity and points of attack/compromise for something that is purely targeted as "something you have" authentication). a couple recent postings somewhat related to threat models and authentication vulnerabilities. http://www.garlic.com/~lynn/2005p.html#25 Hi-tech no panacea for ID theft whoes http://www.garlic.com/~lynn/2005p.html#26 Hi-tech no panacea for ID theft woes --------------------------------------------------------------------- The Cryptography Mailing List Unsubscribe by sending "unsubscribe cryptography" to [EMAIL PROTECTED]
