With the advent of SQL servers, it is not only possible but easier than ever to replicate and share data across the globe. But there are many ways to accomplish this task and choosing the way that is suitable and convenient for you is important.

• Transactional Replication- If you want to provide read only versions of an existing data and send it to separate applications, then you have to replicate the core tables and send them to separate servers. The best thing about this method is that you can replicate as much data as you want and send them to as many servers as required. Also the replication keeps itself synched and in case a problem comes up, you can always reinitialize a subscription to a database. Transactional replication can also be performed on a peer to peer basis. This method is especially useful if you want to make replication changes node by node. But standard transactional replications have some drawbacks in this area. One such drawback may be a definitive slag in replication if the server is down. In this case, re initialization also may take a long time if the file or database is lengthy.

• Messaging Systems- If you want a single central database to feed data to numerous applications, then you should consider messaging systems like Microsoft Biz Talk, Vitria, IBM Web Sphere or TIBCO. These applications have been specifically designed for the messaging system. However these application tend to be expensive and cumbersome both to install and maintain but if you have dozens of applications that all need to share data at some point or a globally distributed system, then these applications will be worth all the cost and hassle.

• Mirrors- Integrating data using mirror system is a good method of data sharing. If you want to make the database accessible to downstream servers or real time, then you can create two asynchronous mirrors where all reads would come from joins to the mirror. The only downside to this approach is that you can only create two asynchronous mirrors at a time. In the case of setting up the mirrors, you also have to use the asynchronous mirrors for the purpose of disaster recovery as well.

• Linked Servers- You can expose the data through linked servers. But this method is can be generally risky so it is advisable to first set up an asynchronous mirror from the central database to another server and then connections of linked server to the mirror. This method will reduce the risks of blocking or performance problems encountered during a query from a web application.

• Bulk ELT Process- If a web application accepts a high degree of latency, then you can write bulk ELT processes which are then moved between servers by the server agents. ELT tools available range from SSIS, Pentaho and Informatica, so you can use whichever works best for you and suits your needs.

    So you can share data across the globe with ease by choosing any of the solutions that suit your business requirements.