mssql database, database synchronization, distributed database, database cluster, how

Because the project needs, the deployment of the product must take into account security solutions and disaster. Because before been doing small projects, basically a single-server, single database structure, but due to a disaster, this issue put on the agenda.

    I qualifications shallow, not very familiar with a lot of things, most recently in the online Baidu a lot of things related to the basic have a rough idea,

    the idea is that
      1 .. do cluster with multiple servers, load balancing pressure to do so first can be split, reduce the pressure on each server, to improve the stability, addition is, a server crashes, another server can continue to run. to provide services.

      2 .. do file synchronization, automatic put some special files on multiple servers do file synchronization folder, such as some pictures resource files, static html. such a server to do the update, other servers can be updated simultaneously, and the user can also upload files synchronized on each server, even if part of a server crash, and ensure that users the use is not affected.

      3 is stumped me, that is a database cluster, or call database synchronization, do not know what the professional scientific name is anti I want to achieve is to function, there are multiple database servers, the content of these databases are identical, access to different sources may be assigned to a different database. When I A1 database, add, modify, delete time All other database synchronization update, when the application is running, it will be resolved to the self-energy minimum pressure on the database server, a database server when hung up when its corresponding program shifts to the other database server, but the key is that the data can not be lost, even if that server's hard disk burned, before the data can not be lost.

      I have found mssql database synchronization features, but there is a big problem is that when two individuals while the two databases when adding the operation would produce two growth from the same id, so two servers clashed concurrent leading to the emergence of inconsistent data. this is a very deadly.

       The above is my description of the problem, I put the posts shared here, a total of future dishes to learn, I hope that the great God give a lot of guidance, presumably these technologies are already many commercial companies use very mature, but there are many people like me dishes not mastered, we interact with it, these days I was so sleepy, I slept, 88.