Best Practices:

• Develop a plan in advance for technical change over time (e.g. what do you do when BIDS 2.0 comes out?)
• Pay off technical debt gradually over time (e.g. regularly check to ensure that any custom software is compatible with the latest version of the language and important libraries).
• Understand the long-term support options for your operating system and analysis software.
• Develop and actively maintain documentation for data and analysis procedures.
• Develop a plan for archiving of unused data to less expensive storage platforms or less expensive tiers of data storage or archiving.

Things to Avoid:

• Waiting until incompatibilities or changes in platforms occur to develop a plan to deal with them

Value Set Definitions: 

• Short: Data collected over relatively short time (e.g., 1-2 years) and no need for active storage post study completion
• Medium: Either data collected over longer time period (3-5 years) and/or need for longer active storage post study completion (e.g., 3-5 years)
• Long: Longitudinal study over many years (e.g., 5+ years) and/or long term active storage post study completion (e.g., 5+ years)

Value of Use Case Example:

Long - It is also possible that they will want to follow these individuals longitudinally and conduct follow-up imaging or behavioral studies and make these data available as well.

Discussion of Use Case:

A potential challenge to the use of cloud computing for long studies is that price and features of cloud computing may change over time, leading to unexpected costs or technical challenges. This is particularly the case if the researcher takes advantage of features that are specific to a particular cloud provider; this can lead to them being locked into that provider.