Victorian Electronic Records Strategy


VERS STORY	STANDARD	ASSESSMENT	PROJECTS	DIGITAL ARCHIVE	TRAINING	TOOLKIT	PUBLICATIONS

Introduction

Standard
(Version 1)
  Specification 1
  Specification 2
  Specification 3

Standard
(Version 2)
  Specification 1
  Specification 2
  Specification 3
  Specification 4
  Specification 5
  Advice 9
  Advice 10
  Advice 11
  Advice 12
  Advice 13
  Advice 14

Home > Standard > Standard (Version 2) > Advice 9 > Loss of records

4.4 Loss of records

4.4.1 The challenge

Records may be lost by system failure. Such failures can include:

Corruption due to failure to accurately copy records from one place to another. This includes errors in copying from one piece of media to another, or from disk into memory. A particular challenge with preventing this type of failure is identifying all those locations in the computer system where records are copied.
Corruption due to failure of indexing. This may result in the records still physically existing but the recordkeeping system 'forgetting' that the record exists.
Hardware failure. Records ultimately have a physical representation, either on media (e.g. disk or tape), or in memory. Hardware failures such as disk crashes can cause the loss of the record.
Disaster. Records will be lost if the computer holding them is damaged by a disaster such as a fire or flood.

4.4.2 VERS approach

The VERS Standard requires the recordkeeping system as a whole to be reliable. In this case the system is not just considered to be the actual recordkeeping application, but includes:

the hardware on which the application runs
the system software, such as the operating system and storage management systems
the processes and procedures that surround the system such as back-up regimes and disaster recovery.

Some failures are due to poor software engineering and can be protected against by ensuring the use of quality software products that have been analysed to identify possible points of failure and have been engineered to guard against them.

Other failures cannot be prevented. It is impossible to prevent hardware failure or a disaster, for example. Such failures must be protected by processes and procedures instituted by the agency and designed to allow recovery of the records. These processes and procedures must be regularly tested to ensure that they work. In many contexts, it would make sense for records disaster planning to be incorporated into the wider organisational disaster-recovery planning.

printer friendly