As we have been discussing, data security wouldn’t be complete without a robust solution to backup your critical information. Often when performing databased backups on complex enterprise document management systems a simple backup of the database does not capture all of the information, data files, metadata and versioning records which may be stored in different locations other than the database. Add to your security checklist backup solutions with sub categories:
- Support for native database backup
- Export formats supported such as XML
- Ability to retrive a single record or file without performing a full restore
When looking at different backup solutions it is important to consider how easy or hard it is to restore a single version of a file. Does the system under consideration require a complete restore to get to a single file?
Data longevity is another major issue for ELN products and their customers. “Data longevity depends on the storage medium and the ability to decipher the information” according to Kurt D. Bollacker who wrote about data longevity for American Scientist in the article “Avoiding a Digital Dark Age.” In the article, the author argues that the issues with Data Longevity is two fold:
- The storage medium used. Tape drives become obsolete, replaced by newer faster versions that are often not backward compatible. Consider the fate of floppy disks, punch cards, etc. These are examples where the data stored on them can no longer be retrieved because the technology, the device to retrieve the data, is no longer available. When considering how best to preserve your research data for 10, 20 and 30 years, consider the devices that will be required to read the medium that you have chosen to use as a back up. Backing up to low costs, slow online hard drives with RAID 1 or another type of replication may be a better solution that backing up to DVD or Tape.
- Ability to decipher the data. The format the data is stored in will need to be supported in the future for the information to be retrievable. Consider if you will need to have the application that created the data to read and retrieve the data. Is the storage format open? Is the data encrypted? Is the data stored with the same extensions and formats that it was created in?
There are many different hardware choices that can be made. The key to remember is that digital storage mediums have expected lifespans of between 5 and 20 years. This means that data validity must be checked regularly, with multiple copies being maintained. The software application should also be treated like the data: keep a master copy, including the supporting OS, and test validate your data integrity regularly.