How to calculate necessary uptime percentage (rules of 9)
Posted on 2014-09-29
I need a philosophical answer. I need to understand how best to determine uptime when it comes to an architecture.
I have learned a couple of things so far:
1. differences in uptime by seconds for each 9: 99% vs. 99.9% vs. 99.99%
2. differences in cost for each 9: $20K vs $200K vs $2M
What I need to understand, is that are there other factors in the rules of 9 other than money lost during downtime?
How do I best determine what uptime I will need when it comes to the importance on the data vs. downtime.
Here's an example. Lets say you have an internal network of servers. These servers have no ties to a sale system, but the data itself is extremely sensitive and important. Prudence dictates that I ensure these systems' data maintains high integrity.
I know integrity and availability are two different concepts, but I am missing the linkage to what type of environment I need based on integrity. Does data integrity and confidentiality drive availability? if so, how? what are the measuring sticks to use for availability to make the best argument for 99.9% vs. 99.99% vs. 99%?