It depends on the requirement of the customer who selects the Data Centers from these two options as these two options cater two requirements sets of Data Centers.
Two data centers geographically dispersed is always preferred for redundancy. Natural disasters, power outages and geo political contingency always should be thought about. Companies that can only afford one location are usually not considered during competitive bidding processes due to risks their potential clients would have to endure. If it’s a consumer only solution without business to business considerations, then you should have a back up agreement with an outsourced vendor In case you experience one of the situations previously mentioned. If you need to discuss further: Michael.t.mclaughlin@verizon.net or 914-907-7705.
This is probably the best answer : geo diversity and redundancy offers a much higher value proposition - it also offers the options to run Primary A and Primary B in Alternance which caters for the top tier of customers requirements
Solution depends significantly on the security requirements and customer business scale, -large businesses would require two-three separated/authotomous Data Centers.
Previous posters have offered solid guidance. I would just add geographic redundancy is twice as expensive but if you are building and selling for a specific customer it might be a requirement regardless of the “tier”. As a side note I know of a company in the United States that is utilizing an old mine for hardened environment. Research BlueBird Networks.
The previous posters have made solid recommendations, especially regarding the benefits of geo redundancy as well as the benefits of potentially having fewer hops to dispersed endpoints. Also, in some locations, there may be other risk factors for a single location beyond natural disaster. One non-USA client required primary and backup due to security and in-country infrastructure limitations. Bottom line - it comes down to the balance between the cost and risk analysis.
Theoretically you may use probability theory in both options. A) Probability of both DCs failing given the uptime in first case B) Probability of both DCs failing given the uptime…then you will have the answer ... Theoretically
Haven't found a solution?