Calculating the Cloud: Determining the True Cost of Hosting Servers in the Cloud (EDUCAUSE Quarterly)

By Adam Goldstein

Creation of a decision tree to evaluate whether a server might be an eligible candidate for infrastructure as a service allows Dartmouth IT administrators to make appropriate cloud sourcing decisions quickly.
Creation of cloud metrics help IT staff determine whether it is more cost effective to host a server in the cloud or keep it in-house.
Applying both tools provides well-reasoned decisions on moving server functions to the cloud or keeping them on premises.
This audio article explains the accompanying poster detailing the decision tree and cloud metrics used at Dartmouth.

Many higher education institutions see considerable appeal in hosting servers and storing data in the cloud. Low cost, ease of use, scalability, minimal infrastructure requirements, and "pay-for-use" make infrastructure as a service (IaaS) offerings such as Amazon Elastic Compute Cloud (EC2) and Rackspace Hosting attractive. However, institutions must consider other factors such as bandwidth limitations, security concerns, legal issues, and service availability requirements when assessing these IaaS offerings.

In an effort to calculate the value of hosting servers in the cloud, Dartmouth College initiated a study to determine the true cost of IaaS offerings. The study was conducted by the Dartmouth Cyber-Security Initiative (CSI), which is an ongoing collaboration between faculty, staff, and students and focuses on projects aimed at improving the security of the college's information systems.

The study resulted in two useful decision-making tools:

Decision Tree: Eight questions quickly allow Dartmouth IT administrators to determine whether a server might be an eligible candidate for IaaS.
Cloud Metrics: For those servers which might be good candidates for the cloud, a series of metrics including bandwidth, storage capacity, processing power, and usage patterns help IT staff determine whether it is more cost effective to host a server in the cloud or keep it in-house.

The audio file below explains the pdf poster of the decision tree and cloud metrics we use at Dartmouth to make decisions on moving servers to the cloud or keeping them on premises.