Abstract

INTRODUCTION Current databases need logging and recovery to maintain the correctness properties of transactions and consistency of the database under failures. Also, logging and recovery have tremendous performance implications. While there has been a lot of work that deals with logging and recovery algorithms for traditional disk resident databases ([1, 2, 3, 4, 5]) and Main Memory Databases (MMDB) ([6, 7]), researchers have not systematically explored the issue of logging and recovery in Real-Time Databases (RTDB). In fact, there is a need 109 110 REAL-TIME DATABASE SYSTEMS: ARCHITECTURE AND TECHNIQUES for designing new algorithms for logging and recovery in RTDBs because the sequential nature of logging and the lack of time and priority cognizance during recovery are not in tune with the priority oriented and preemptive nature of activities in RTDBs. This chapter motivates the need to do logging and recovery differently in Real-Time Databas

Citations