How is data on DataShare preserved?
Answer
Materials in DataShare are monitored and preserved at the bit-level. Bit-level preservation ensures that files stay intact. More advanced preservation techniques such as reformatting and migration, are not offered at this time. For this reason we encourage all deposits to be made in nonproprietary and/or ubiquitous formats whenever possible.
Data published on DataShare is kept for a minimum of 5 years. After five years data sets will undergo retention reviews to determine their continued relevancy and usability. No data will be removed without first attempting to contact the depositor or PI. If your research requires a longer minimum retention period please let us know before the data is published.
More about bit-level preservation: DataShare's data is stored on Amazon AWS S3 (Simple Storage Solution) infrastructure. S3 is designed for 99.999999999% durability (i.e. file integrity) and keeps redundant copies in case the primary fails. Additionally each file is given an MD5 checksum when uploaded. File integrity is regularly checked against the checksum to make sure the file is intact.