Decision intelligence for data center disaster planning
Much of the world’s information is handled in data centers: nondescript on the outside, indoors you’ll find a din of humming machine in acres of racks. If an earthquake or other natural disaster should strike a data center, the cost can be enormous.
So spending money on risk mitigation strategies like earthquake-proofing makes sense.
But it’s a tricky balancing act: spend too much for a disaster that never happens, and it’s money down the drain. Spend too little, and a disaster can spell tremendous costs.
A decision model can help. We built a demonstration of one shown in the video above, at the link here, and in the bottom of this article, for those viewing this online.
Give it a try, it takes about one minute:
- Choose a data center location by clicking on one of the “select” links below the graphic.
- Think about what percent of customers you think you’ll lose for a data center outage. Use that to set this assumption slider that looks like this:
- Think about how long it’ll be until your data center comes back on line after an outage. Set that amount on this slider:
- Now, experiment with the lever slider reflecting different amounts of investment in risk mitigation:
Now, observe as you do this that 1) the graphs at the right change, reflecting the expected cost of net churn, SLA violations, and credits. Also observe that 2) it takes a while for the graph to equilibrate. What’s happening here: a complex dynamic system feedback loop is running behind the scenes: an investment in risk mitigation shifts the risk bottleneck, which changes the impact of further risk dollars.
This simple model (along with others) shows how decision intelligence is used in practice. In particular it shows how levers, assumptions about externals, and a complex system model interact to support important decisions that drive high-value outcomes.
Again, if this is hard to read and/or truncated, click here to view it on its own page.