在MapReduce中，reduce任务与reducer有何不同(In MapReduce, how does reduce task differ from reducer)

 对于ID的问题，您应该使用GUID作为ID，甚至不要使用身份主键尝试此ID。 这允许您在客户端，Web服务器等上创建ID，并且它们不会重复。 
 
 对于同步问题，有不同的工具和第三方应用程序将为您执行此操作，但我不建议您使用您的程序，因为您可能有1000个客户端都需要复杂/昂贵的同步工具。 您应该编写一个可以发送/获取更改的Web服务，并在表中添加DateTime字段，以便在记录已存在时告知哪个数据库具有最新数据。 这需要一点点工作，但这就是我要做的。 
 
 顺便说一句，如果您使用的是VS2015而您的客户端都是Windows 10，那么有一些内置功能可用于构建为您同步的应用程序。 但是我们可以期待用户拥有Windows 10还需要一段时间。 

For the issue of ID, you should use a GUID for the ID, do not even attempt this with an identity primary key. This allows you to create the ID on the client, web server, etc and they won't duplicate.
 
For the issue of synchronization, there are different tools and 3rd party applications that will do this for you but I wouldn't suggest doing that with your program because you may have 1000's of clients that would all need complex/expensive synchronization tools. You should write an Web Service that you can send/get changes from and add a DateTime field to your tables so you can tell which database has the latest data when the record already exists. It takes a little work but this is how I would do it.
 
BTW, if you are using VS2015 and your clients are all Windows 10, there are some build in features for building apps that synchronize for you. But it will be a while before we can expect users to have Windows 10.