Is there a central DB SERVER in this kind of situation?

I have multiple servers containing the same table schema, but each responsible for a portion of the overall data. No o开发者_Python百科ne machine needs to contain all the data. Each machine only contains data relating to its department.

Is there a central database server which will contain all the data, or just a central database server which can be used to access the other servers?

---

It sounds like what you are doing is partitioning the data. Remember the CAP Theorem states you cannot have accessibility and data consistency all the time in this configuration. Normally you'd have data servers (like in your example) and then different servers for read/writing using something like SQL. They can be distributed (just as long as they know about all the data servers). Just remember that 'in a distributed system there is no central arbiter of truth'.

We use the same came of setup in a production network, where the faster servers are a cluster with the last 2 years worth of data and then there are slower archive servers. Results can span the entire collection of servers depending on the query. If a report needs archived data it queries a separate server that performs a join behind the scenes.