Optimizing Economics of NoSQL and Big Data in the Public Cloud

Doing Big Data and NoSQL in the public cloud sounds great, but at some point, someone has to write a check to make it happen. Is there a strong financial argument for making this move? Can you start small and grow larger easily? This paper answers this question with a detailed Total Cost of Ownership (TCO) analysis for open source data management in the public cloud. Contrasting cloud deployments on the standard vs hyperscale CenturyLink Cloud stack Powered by Intel Cloud Technology, versus the on-premise analogue, the paper models the relative financial merits of each approach. The paper considers the best potential use cases for on-premise versus cloud and walks through an example use case to show the impact to business outcomes (time to results, cost, TCO).

Cost factors compared include:

– Time to deploy: on premise (Existing HW vs HW purchase and physical provisioning) vs public cloud (std vs hyperscale) > business impact (EU / IT)
– Data center overhead vs. public cloud costs, considering

• Power consumption
• “Lights out” management of hardware
• Security and malware protection at the hardware and software levels of the stack

– Costs of scaling the deployment in two different scale add-on scenarios
– Costs of system administrative personnel
– Competitive Impact of faster/more accurate results