Smart-Card Transactions

Transactions processed by smart cards start with the selection of an application and the reading of the corresponding data on the card's EEPROM. Then the application identifies itself so that the terminal knows which application is to be run. A transmitted application-file locator tells the terminal which files and records to read. If the application contains a public-key certificate, the terminal verifies the public key through static authentication.

Then the terminal decides whether a transaction should be processed with the card issuer's host computer on-line or is safe enough to continue off-line. This decision is made by the terminal's built-in risk management system. The t erminal then requests either an on-line or an of f-line authenticator from the card. At the same time, the card makes its own risk evaluation and selects either an off-line or on-line mode.

The card can overrule the terminal's decision to process off-line, but it cannot overrule an on-line transaction. If the card returns an off-line authenticator, as requested by the terminal, the transaction is successfully completed. Otherwise, the terminal contacts the card issuer's host computer, which returns a cryptogram to check the card's authentication and transmit the data.

The OSes of older smart cards often cause problems when the card is pulled out of the reader before the session is completed. Newer OSes organize data transmission as a series of database transactions. If a transaction is aborted before completion, the card rolls back all the data to its original state. If a session is aborted during the transmission of the data-reception proof, the terminal resends t he proof until acceptance, or the card deletes the update. Thus, users can't lose electronic money due to an interrupted session.