| As the high-speed development of digital information and Internet technology, we have entered the period of "Pervasive Computing". Embedded database technology as one of the most influential technology in the future has become the focus of the database. Meanwhile, with the development of embedded Internet and the demand of dealing with a great deal of data, the software system in electric power meets new problems, the security,real-timing of data processing. In this paper, as the deficiencies of security in the traditional embedded real-time database, we design a general framework of the secure embedded database management system SEDBMS, It is suitable for the electric power system, and also realize the audit subsystem.In the design of the general framework of SEDBMS, we first introduce the characteristics of the embedded database and the principles of design, then analyze the demand for the electric power system, finally the overall framework SEDBMS is to be designed focusing on security and real-timing.From the aspect of real-timing, according to the embedded database's real-timing characteristic, the conventional transactioç–¤?heduling methods can not real-time dynami??y schedule the different types of transactions, which can not deal with real-time transactions timely and complete critical real-time transaction on time to meet the electric power system. We propose a new dynamic priority scheduling algorithm to meet transaction demand for electric power system, and simulation results show the effectiveness in ensuring the key transactions.From the aspect of security, we select the suitable security policy according to the electric power system, redact the traditional effective audit mechanism, design and realize a new little audit subsystem for embedded database system, which is the last security line of defense but also ensure the real-timing of database system. It effectively prevents the attacks from inside and outside of system, and improves the security of the system.Finally, we design test program to test the audit-based intrusion detection subsystem, and according to test results to analyze the performance of the system. |