1. Introduction How to achieve data synchronization in different regions in the development of cross-regional storage management systems, how to maintain data consistency, and become the key to system success or failure. We successfully solved this problem, and now the following is as follows. System Environmental Database System: SQL Server 2000 Operating System: Windows 2000 Advance Server, IIS 4.0 Network: The Sales Center is connected to the Internet through a special line, and some regional points are connected to the Internet. 2. Copy policies consider various factors, such as the tightness of the data connection, the reliability of the network, the cost, etc., we choose to disperse processing data, timing synchronize data. To do this, we divide the data into two categories: only one-way copy data (such as the basic data of the system, a part of daily processing data), data for two-way copying (part of daily processing). Simply a one-way copy of data can be divided into: Maintenance by the sales center, distribute data to each split point; maintained by each split point, summarize data to the sales center. Different types of data need to be different replication strategies. Maintenance by the sales center, distributing data to each of the points to set itself on the SQL Server of the Sales Center as a publishing server and distribution server, defining publications, using things to copy models and anonymous subscription methods. Set the pull subscription on the SQL Server of the distribution point. Maintained by each split point, summarize the database of the sales center. Set its own SQL Server as a publishing server and distribution server, SQL Server in the SQL Server in the distribution point as a subscriber. Define publications and use things to replicate models. And set the push subscription, copy the data to the sales center.
Data that requires two-way copying sets its own as a publishing server and distribution server on the SQL Server of the Sales Center, defining publications, using merge replication models and anonymous subscription methods. Set the pull subscription on the SQL Server of the distribution point.
Why use the above replication strategy? This is because there is a fixed IP on the computer of the sales center, while only dynamic IPs and data have different categories on the computer of the distribution point.
The above is only a brief introduction to a solution, and you are looking for more valuable comments.
