This is part of Business Continuity Planning (BCP) where natural disasters can bring an end to the Hadoop system, if not planned correctly.
Here, the strategy would be to use multiple geographical regions as providers to run the big data systems. When we talk about multiple data centers, the obvious challenge is the network and the cost associated with managing both systems. One of the biggest challenges is how to keep multiple regions in sync.
One possible solution is to build a fully redundant Hadoop cluster in other geographical regions and keep the data in sync, periodically. In the case of any disaster/breakdown of one region, our businesses won't come to halt as we can smoothly run our operations.
