Distributed Database Log Management System

Along with the progress of technologies in computer networks, distributed database has undergone significant development and become one of the hottest research fields. Among these techniques, log system is the foundation of fault-tolerance in distributed database systems. Many sub-systems can benefit from powerful log system, such as distributed query, parallel processing control, transaction recovery, and message-based fault-tolerable log techniques. In addition, log can be used to maintain consistency among database's duplicates. Log has become the most popular and dependable technique to maintain system reliability.However, little research has been carried on in the field of log manager system in distributed databases. As a result, few manage tools have been developed in this field. Even worse, many distributed database system don't provide log-based recovery mechanism. The main reason is the confliction among clients' different requirements. Based on a distributed database DPSQL, the project try to provide a quality log manager system. First, this article will analyze the failure model and point out the necessity to employ a log manager system. At the same time, we also discussed many most important features of log system in distributed database system. Then, the design goal of our log manager system is presented. Following this goal, we discussed the system model, design details, and performance feature. This article aims at how to enhance the log system's performance. After analysis of the factors affecting system performance, we will discuss the disadvantages of conventional log system. Then, the article will emphasize on the details of the WAL rules, check-pointing techniques, buffer mechanism and write-through rules in DPSQL.The article presents a unique log structure based on logical view of database. This log structure can satisfy at-least-one syntax and can be executed at different places. It also supports multiple recoveries. Besides the efficiency it provided to log manager system during failure-free time, this log mechanism will considerably improve, the performance, of transaction process and log-recovery protocols.Besides database's integrity and consistency, the log manager system also obtains the high performance during failure-free time at the price of acceptable overload of disk I/O during recovery.