Research And Implementation On Performance And Architecture Optimization Of Finacial-Oriented Java Card

Java Card is a smart card which java applets can be run securely on. With the global EMV migration and vigorously promote of Golden Card Project, the application of Java Card is more and more extensively. The Java Card replacing the issued bank cards is now imminent. However, the problem which caused bad runtime properties and system components which can’t be updated, is still hampering its development since Java card’s emerging. This paper is wrote to research the architecture of software and hardware and operating principle of Java Card, and to optimize the execution engine’s implement and memory system infrastructure, to get breakthrough in promoting the technical development of Java Card.This paper proposes the feed-back JCVM instruction prescheduling scheme, which can collect runtime execution flow of bytecode instruction and native method on financial field application, and then we proposes the concept of Weighted Control Flow Graph, and one kind of instruction arrangement algorithm based on it, which realizes the prescheduling of interpreter and native method module. In the target system, we reorder the hot instruction handler functions and native methords according to the statistical information analysis of the feedback application, and increases the runtime locality of Java Card execution engine; It proposes a MMU based memory architecture design of Java Card, and a scheme of system components update of Java Card, so that it achieves the Java Card system component update without issuing a new card and the destruction of the safety to deal with the emergence of the system vulnerability.Finally, through experiment validated, the paper analyzes the performance improvement by the scheme of feed-back JCVM instruction prescheduling, and the feasibility of the scheme of the MMU based system component update of Java Card. Based on the contrast test on Java Card performance evaluation system between our project system and some typical commercial cards’, put forward the further optimization suggestions.