I can't understand the difference between transactional replication and merge replication. This is my scenario: In an organization I have a SQL server which need to collect information from different sql servers which are located in different parts of organization or around the city and some report will create according to gathered information. Data in different SQL servers update every 5 or 6 minutes. I don't know should I use transactional or merge replication?
Transactional replication is more like real-time, but you have to have a constant or near constant connection between the subscriber databases and the publisher databases. Changes are made to the publisher database only and then those changes are pushed out to the subscriber databases.
Merge replication is different. Both databases can be updated independently, then, at scheduled intervals, the data is compared, and merged. If there are inconsistencies, those are resolved depending on the algorithm you set when you set up the replication. It's slower, but better for applications that aren't centrally hosted.