Too often systems are built with the faulty expectation that they will last forever. In reality, systems go through a life cycle, which ends in their complete redesign or retirement from service. The need or requirement to retain accessible and useable electronic records may exceed the life of the system that created them. Electronic records created by one system may need to be moved or migrated to another system. System migrations are extremely complex and should be planned for and accomplished before the original system becomes obsolete and inoperable. Migration should be implemented incrementally along with periodic system and software upgrades and should include quality control checks. While migration has become common, it is still fraught with danger. For example in one case involving FDA-mandated records of drug testing, blood pressure numbers were randomly off by up to 8 digits following data transfer from UNIX platforms to Windows NT operating systems.⁸

The least complex form of migration is simple data migration where the data is pumped from the old system into the new system. For low risk electronic records this may be sufficient to retain them in a useable form. However, even such a seemingly simple task could be problematic depending on the complexity of data structures and the use of proprietary formats. Furthermore, a successful migration of high-risk records will require that information in addition to the informational content of the records be migrated to ensure their integrity and reliability. Information relative to the electronic record’s creation and use such as metadata, audit trails, authoritative controls, and documentation need to be migrated to the new system and maintained for the same retention period as the records. In other words, it is not enough that the content or data of the records be migrated to the new system. The context in which the records were created and their structure needs to be maintained for the life of the records as well. The migration of this additional information could be extremely difficult and will involve additional planning and resources.

Most installed technology involves proprietary systems and formats. Proprietary data formats can greatly complicate migrations and jeopardize the accessibility of electronic records. Technology policies should strive to establish standard formats for electronic records. Since software is subject to change – either by the implementation of new releases, by changes to operating systems, or changes in hardware requirements, the use of non-proprietary formats is strongly recommended. Regardless of the medium on which a record is stored the use of non-proprietary formats will minimize the long-term costs associated with maintaining the reliability of and migrating records. The use of widely adopted standard formats (relational databases, ASCII, SGML, etc.) can help reduce the rate of technological obsolescence and the frequency of migrations, as well as facilitate migrations. Be aware, however, that standards change or are replaced over time and must be monitored. The National Institute of Standards and Technology (NIST) is exploring the use of standard e-records storage formats.

Although not a permanent solution, migration is the primary solution for retaining electronic records over extended periods of time, especially if there is a need to retain the records’ original functionality. However, other possible solutions to long-term retention are also being explored including:

• Encapsulation: Encapsulation refers to a method of capturing the look and feel of the original record along with any required metadata as a single digital object in a portable format. In some ways, encapsulation combines system migration with use of standard formats. Encapsulation strategies are just beginning to be investigated.
• Emulating obsolete technology: Emulation consists of using hardware and software to allow one computer technology to act as if it were another technology. This solution allows e-records to remain in their original file formats while the hardware and software change. Emulation is complicated and expensive to achieve for any sophisticated system. Research on emulation solutions is ongoing.

If loss of a record series would place an agency at significant risk, exporting the records to a technologically neutral, durable media such as computer output microfilm or paper as insurance against unforeseen migration problems. Because export will result in a loss of system functionality, this option is unattractive and clearly reserved only for records of extraordinary value. In some instances, it may be possible to export a subset of the essential information. A hybrid approach that preserves the records in both electronic and durable formats can offer functionality and confidence of preservation.

Export to physical media requires the preservation of sufficient context and structure to ensure that the records are acceptable as evidence. This information may be appended as a header or footer, although some can be translated back into the media’s physical characteristics. For example, a message digest used to demonstrate the record’s integrity is of no further value because the content is fixed on film or paper and subject to traditional forensics tests for altered documents. However, the process of exporting records should verify the message digest as part of the export process so that the records can be certified as authentic.