In this section
Illustration by Jørgen Stamp digitalbevaring.dk CC BY 2.5 Denmark
This section together with Why digital preservation matters, is designed as a briefing for those new to digital preservation. It is structured into three inter-linked sub-sections covering Threats to digital materials, Organisational issues, and Resourcing issues. It links to more detailed treatment in other sections of the Handbook as appropriate, but has a particularly close relationship to the Getting started section, which is also particularly designed with those new to digital preservation in mind.
Digital preservation can often seem daunting at first. It is important to realise than those with existing skills in either information management or information technology within organisations are well placed to build on and apply these skills to digital preservation activities. However, it may require initially learning some new unfamiliar terminology (see Glossary), extending skill sets, and sometimes working in new ways.
Threats to digital materials
Keeping the data
Every digital file is formed from a series of zeros and ones, or bits (binary digits). These streams of bits need to be captured and retained over time, without loss or damage, to ensure the survival of digital materials. There are an array of threats to any attempt at preserving these bits. Storage media can decay over time, leading to corrupted files. Storage media may become obsolete and unsupported by contemporary computers and the software that understands and provides access to them. The bits may be ignored, abandoned, accidentally deleted or maliciously destroyed. Removable media could be left on a shelf and forgotten, files stored on a shared network drive might be left without an owner, or a third party cloud storage provider could go out of business.
Maintaining a systematic process for bit preservation remains a fundamental requirement in ensuring long term digital preservation. Storage media must be monitored and refreshed (See Legacy media). Redundancy must be introduced by replicating or backing up files, introducing diversity in dependent technologies and avoiding catastrophic disaster at a single geographical location (see Storage). Checksums must be generated and frequently recalculated to identify any loss and ensure that the integrity of the bits can be verified in an efficient and automated manner (see Fixity and checksums). The locations in which digital materials are stored should be carefully recorded, and responsibility for their preservation allocated.
Keeping the meaning of the data
Reconstructing the information that is encoded within a stream of a bits typically requires computer software that is designed to render, manipulate, analyse or otherwise interact with the particular encoding or format of the data. Over time, the encodings (or file formats) may change, and the software applications that interact with them may go in and out of favour. Although unusual for well known file formats, less well used file formats may become obsolete over time, as the software that renders them is no longer supported (see File formats and standards).
Understanding the technology on which particular digital materials are dependent enables appropriate action to be taken to ensure their preservation. A considered preservation planning process might result in the migration of digital files from format to format, the emulation of obsolete software, or the employment of alternative software applications to render the data (see Preservation action). Each of the options presents its own advantages and disadvantages and these need to be evaluated carefully, possibly on a case by case basis (see Preservation planning).
While file format obsolescence has not emerged as the overwhelming danger that was previously perceived, challenging subtleties remain. It may be possible to find a method for rendering an old file format (perhaps by emulating some obsolete software), but how accurate is the rendering, is it legal to run the software, and how much will this complex effort cost the preserver and the user?
Maintaining trust in the data
Digital materials have the potential to remain fluid over time, being edited or altered with ease, being damaged by media failure, or decoded into human readable information in an unreliable or inaccurate manner by rendering software. For an end user to have trust in the result of digital preservation work it requires careful consideration of the entire lifecycle of the digital materials and who or what has interacted with them over time. Information management systems need to be able to link to essential contextual information regarding the business procedures of the creating agency. Authenticity and integrity of digital resources can be equally important in other sectors. For example, scholars will need to feel confident that references they cite will stay the same over time, courts of law will need to be assured that material can withstand legal evidential requirements, government departments may well have legally enforceable requirements regarding authenticity, and so on. This issue overlaps with both legal and organisational issues and it may be one which is best resolved within individual sectors rather than through generic procedures.
The application of data integrity techniques and the maintenance of audit trails can provide confidence that a digital object has remained unchanged (except by necessary preservation action) since deposit in an archive (see Fixity and checksums, and Information security). Ultimately its authenticity to a user may depend much more on the broader trustworthiness of the preserving organisation as a whole. Maintaining high quality preservation processes based on current best practice and validated by appropriate audit and certification will be crucial (see Audit and certification).
Keeping the context of the data and it's dependencies
The meaning of digital information can be dependent on additional information that may have been implicit within the context it was originally created or used in, but less clear when revisited at a later date. Identifying, understanding and capturing relevant contextual information can be vital to a successful preservation effort. This might be as simple as capturing the units of measurement used within a spreadsheet, the scale of a map, or the point of origin within a CAD drawing. As digital information continues to be created in a more complex and interconnected manner, it may be necessary to retain the place of particular digital materials within a wider context of associated information resources. What may be seemingly simple and stand alone documents may actually depend on related files, referenced fonts and may have pointers to related information on the web. What might be viewed as a simple web page may have been generated on the fly from live data sourced from different locations on the Internet.
Understanding the data, how it will be used, its dependencies and its context will enable it to be captured for preservation in an appropriate manner and documented in a sufficiently explicit manner to enable the intellectual content to be retained and understood on into the future (see Metadata and documentation).
Acting in a timely manner
Prioritising digital preservation activities and applying them in a timely manner can be crucial not just in avoiding loss but in ensuring the best use of limited resources. Where the opportunity exists to intervene early in the lifecycle, digital materials can be shaped to survive better into the future. The choice of file format, the capture of critical documentation or the description of key relationships in the metadata may require a small investment up front, but could deliver considerable savings further down the line (see Creating digital materials). Where this is not possible, and risks to the data have been identified, the best timing for preservation action can be unclear. Early intervention to head off technological obsolescence may provide greater confidence of long term sustainability but with the risk that intervention may not ultimately be necessary and resources were wasted. Just in time action may minimise unnecessary activity, but increase the effort needed to research obsolete technology in a particular case requiring specialist knowledge that is no longer current. Appropriate action should be taken on a case by case basis.
Coping with the data deluge
Research reported by David Rosenthal noted that the rate of data creation is expanding by about 60% per annum; that developments in data storage allow are expanding at about 25% per annum; and that data centre budgets are expanding at about 2% per annum (Rosenthal, 2014). While this places challenging pressures on selection policies and other organisational decision making it also poses technological questions. Simple preservation processes that function effectively at one level will not necessarily scale easily to work with very large volumes of data or perhaps very large individual files. The technology and understanding to work at scale is moving forward rapidly, with growing expertise for handling large audio visual collections, research data and web based archives (see Content specific preservation). But some repositories still face significant challenges in developing and maintaining scalable architectures and procedures to handle growing quantities of data. The technical and managerial challenges in accessioning, managing and providing access to digital materials on this scale should not be underestimated. It can be important to remember that selection, appraisal and disposal are significant components in any digital management activity.
While technological issues can be challenging, there are also numerous challenges which relate to organisational issues. These include how digital preservation is organised and delivered, or how those responsibilities change over both time and the lifecycle of digital materials. There are common digital preservation challenges faced across organisations, yet every organisational context will be different. It is vital to ascertain organisational drivers and tailor practical solutions to meet these needs. There is no one size fits all approach for digital preservation.
The creation, preservation and access for digital materials are widely distributed. As a result, there is an increasing need to go beyond the confines of individual organisations, or even countries, to maximise the benefits of the technology, address common issues, and to overcome the challenges cost-effectively.
In-house or outsource?
The decision whether to do all or part of digital preservation via a third-party or in-house, or perhaps a combination of the two, is often a complex one. Digital preservation may be undertaken in-house if there is sufficient staffing and infrastructure but outsourcing some activities or support can be cost-effective, and can leverage internal capabilities and capacity.
Outsourcing specific tasks or services from a repository is by no means a new phenomenon. Repositories have contracted out some of their operations for decades. Of critical importance is having and retaining sufficient knowledge to be able to prepare effective specifications and monitor performance. Outsourced work must be easily verified and quality checked, and this is best enabled via careful design of the specification, and the reporting providing by the 3rd party. Cost will clearly be a key consideration when deciding whether or not to contract out digital preservation but there are also other factors to consider such as legal issues. For example, legal provisions due to privacy or confidentiality may influence whether outsourcing is appropriate or not. The advantages and disadvantages of each option will need to be balanced in light of the individual organisation's mission and responsibilities (see Procurement and third-party services and Cloud services).
There is a significant overlap in the digital preservation issues being faced by all organisations and across all sectors so it makes sense to pool expertise and experience. There are compelling reasons and, in some cases, political pressure, to engage in greater collaboration within and between organisations in order effectively to confront and overcome the challenges of digital preservation.
Most organisations readily acknowledge the benefits of increased collaboration but also indicate the potential difficulties that can arise in the form of differing agendas, timescales, or funding mechanisms. None the less, it is often possible to collaborate in specific areas or with different levels of intensity that moderate these potential difficulties. Some of the most high-profile and successful initiatives in digital preservation of recent times have been collaborative in nature (see Collaboration).
The modern digital world is a place of both rapid technological and organisational changes. Organisations re-organise internally, merge, or cease to operate with increasing frequency. Digital preservation is a long-term activity and the likelihood of it being affected by organisational change increases over time. This may affect a repository not only through changes to its parent organisation, but through changes to its major depositors and users, suppliers, or collaborators. Organisational change is therefore a major risk to be managed (see Risk and change management).
The nature of the technology and dependencies in the preservation of digital materials are such that there are implications for organisational structures. Many of the activities converge, for example decisions about acquisition and preservation should sensibly be made at the same time. Organisational structures will need to cross boundaries in order to draw on the full range of skills and expertise required for digital materials. Assigning responsibility for preservation of digital materials acquired and/or created by an organisation will inevitably require involvement with personnel from different parts of the organisation working together. This can potentially present difficulties unless underpinned by a strong corporate vision which can be communicated to staff (see Collaboration, Advocacy, and Staff training and development).
Roles and responsibilities
There are some existing repositories which undertake responsibility for specific subject areas or specific formats. In the UK, for example, the UK Data Service undertakes responsibility for selected social science research data, while the British Library's National Sound Archive assumes responsibility for its collection of sound recordings. Each repository will need to consider its own collection policy and the broader landscape of collecting institutions and remits within which it sits.
The digital environment demands engagement with a large group of stakeholders. The lifecycle approach to digital preservation advocated in the Handbook has significant implications for the way organisations responsible for long-term preservation need to interact and collaborate with creators, publishers and other intermediaries, and each other.
Creators of digital materials need to be able to understand the implications of their actions in terms of the medium to long-term viability of the digital material they create. Whether it be a record created during the day-to-day business of the department, a digital copy of analogue collection material, or a "born digital" resource, guidance and support as well as an appropriate technical and organisational infrastructure will assist in facilitating greatly improved prospects for efficient management and preservation (see Creating digital materials).
The enormous quantity of information being produced digitally, its variable quality, and the resource constraints on those taking responsibility to preserve long-term access, makes selectivity inevitable if the objective is to preserve ongoing access.
In the digital environment non-selection for preservation may almost certainly mean loss of the item, even if it is subsequently considered to be worthwhile.
In cases where there may be multiple versions, decisions must be made in selecting which version is the best one for preservation, or whether more than one should be selected. Sampling dynamic resources as opposed to attempting to save each change, may be the only practical option but may have severe repercussions if the sampling is not undertaken within a well-defined framework and with due regard to the anticipated contemporary and future needs of the users.
Some consideration also needs to be given in the selection to the level of redundancy needed to ensure digital preservation. There needs to be a clear understanding of who will undertake that responsibility and for what period of time. Otherwise, even if several copies are stored in various repositories, all of those repositories might, for a variety of reasons, cease maintenance of the digital object at some point (see also Acquisition and appraisal).
Balancing security and access
There has always been a strong link between preservation and access. Repositories need to ensure that their digital materials are safe and secure, but most also provide access to a variety of users. Access by real users can provide a valuable steer to the design of preservation facilities, helping to avoid unnecessary actions but also validating and introducing a feedback cycle.
Many types of digital material selected for long-term preservation may contain confidential and sensitive information that must be protected to ensure they are not accessed by non-authorised users. In other cases there may be legal or regulatory obligations on the repository affecting access. There can be tensions between these two roles and a need to strike a balance between security and ease of access (see Access, and Information security).
Legal issues are not simple in digital preservation. Multiple copies and derivative versions often exist of digital materials, and there may be associated software and metadata with them from different sources. Digital content is generated by a wider group of creators and incorporates more diverse formats and intellectual property rights (IPR) than applies in the analogue world. The law also often lags behind technological change and digital preservation needs. Some of the key legal issues that affect repositories in collecting, preserving, and providing access to digital materials are:
- Any legal requirements in terms of management, preservation, and access placed upon the repository and its parent organisation, by donors and funders via contracts and agreements or via legislation by Government (e.g. accessibility, availability, information security, retention, audit and compliance, Public Records, Legal Deposit, etc.);
- Those legal obligations relating to third party rights in, or over, the digital materials held by the repository (e.g. copyright, data protection); and
- The legal elements of any relationship between a repository and any third-party provider or providers (e.g. terms of service contracts and service level agreements).
Budgets and costs
The cost of digital preservation cannot be easily isolated from other organisational expenses, nor should it be. Digital preservation is essentially about preserving access over time and therefore the costs for all parts of the digital life cycle are relevant. In that context even the costs of creating digital materials are integral in so far as they may need to include cost elements which will ultimately facilitate their long-term preservation (see Creating digital materials).
The ability to employ and develop staff with appropriate skills is made more difficult by the speed of technological change and the range of skills needed. It is also limited by resource constraints on organisations which may well need to manage growing traditional collections and digital collections without additional resources.
Nonetheless the exercise of calculating costs, however complex, is a valuable and necessary task to establish cost-effective practises and a reliable business model. The cost of the labour required for digital preservation will be the most significant by far and includes not only dedicated experts but varying proportions of effort from many staff such as administration, management, IT support, legal advisers etc.
Other major issues to impact costs include organisational mission and goals, including the type and size of collections, the level of preservation committed to, the quantity and level of access required, and time frame proposed for action. These are discussed in detail in the section on Business Cases, benefits, costs, and impact..
The relationship of costs and institutional strategies and activities such as Collaboration, Procurement and third party services, Legal compliance, Staff training and development, or Standards and best practice are also discussed in the relevant sections of the Handbook.
Staffing and skills
Digital preservation involves a range of skills and organisational roles. Typically digital preservation draws on a range of skills which are not normally found in combination. That means larger organisations will likely need to assemble multi-disciplinary teams while in smaller organisations it will be necessary to rely on a distributed team or sources of support.
There are three main issues to consider with respect to staffing and skills:
- Firstly, although there have been considerable improvements in recent years, digital preservation teaching often lags behind current best practice or is wholly theoretical within relevant information management programmes for new entrants into the profession. So individuals with practical skills and experience are in high demand and staff can be hard to recruit.
- Secondly, job descriptions can be hard to script, especially when agencies are effectively starting from scratch with a new role. To this end a number of research projects have attempted to describe generic skills needed for digital preservation, using as a basis the assumption that different skills are required at different levels of an organisation. Tools like the DigCurv Skills framework allied to the Digital Preservation Coalition's Vacancies section can be very useful when describing new roles. Larger organisations with multi-disciplinary teams may be able to recruit to roles that are 'digital' variants of existing professional categories such as archivist, librarian or records manager, but for most organisations new types of roles must be created.
- Finally, staff working in digital preservation frequently report the need to engage in active career development. Given the expectation that technology and the needs of users develop through time, so the staff involved in meeting these changing requirements will need to find ways to have their skills constantly refreshed, such as through specialist briefings and professional networking (see Staff training and development).
Effective digital preservation requires some basic facilities or infrastructure, typically technological in nature, on which operational workflows and the processing of digital material can be based. While these may be rudimentary or at least small scale in nature when an organisation takes its first steps in digital preservation, ramping up operations to address large quantities of data will require considerable investment in the facilities required to support it.
With the typical requirement of replicating preserved data to avoid loss, storage hardware remains amongst the most important digital preservation facilities. Storage technology has changed rapidly over recent decades. Archives widely used media such as CDs or DVDs for long term storage, but the rapid developments in magnetic media have brought fast and reliable storage that has made handheld media redundant. Enterprise storage systems now provide large storage volumes at reasonable cost. While they have finite lifespans, typically of around 4-8 years, they are easy to monitor and then replace when they reach end of life (see Storage).
Organisations may also wish to consider cloud services to "rent" preservation infrastructure. The flexibility of the cloud allows relatively rapid and low-cost testing and piloting. Cloud services can provide easy, automated replication to multiple locations and access to professionally managed digital storage and integrity checking. Repositories can add access to dedicated tools, procedures, workflow and service agreements, providing a digital repository system tailored for digital preservation requirements via specialist vendors (see Cloud services).
Digital repository systems
Many of the core requirements for preserving digital materials are provided in an automated fashion by dedicated digital preservation systems, or trusted digital repositories. A repository application will uniquely identify each digital object placed within it. It will manage the storage of that object, identify its characteristics and help a repository manager to plan its preservation. It will also facilitate access to the object. While basic preservation can be provided on an ad hoc basis at a small scale, a dedicated repository application is essential to managing digital materials effectively over time. The OAIS model provides a high level model for the functions required by a repository (see Audit and certification for more information on certification of trusted digital repositories and Tools for repository systems and components).
High performance computing
Increasing volumes of data require not only more storage but also greater computational power. Characterising and assessing the technical characteristics of data, indexing data to enable search and access, integrity checking and a host of other tasks require considerable computational performance. Those dealing with these big data, be it research data or web archives have typically looked to high performance computing, and technologies such as Apache Hadoop running on clusters of commodity hardware to meet this need.
Digital preservation laboratory
A number of larger organisations have developed lab environments within which an array of old and new technology can be applied for the stabilisation or ripping of data from obsolete media, and has been championed by organisations working with personal digital collections. Specialist drives for reading magnetic media, robots for processing large numbers of optical disks and write blockers for allowing access to hard drives without changing the bits in the process, are just some of the equipment that could be useful here. Media recovery companies offer an alternative approach that may be preferable in high volume cases, albeit with less control of the process and the need to move media offsite (see Digital forensics).
How Toy Story 2 Almost Got Deleted: Stories From Pixar Animation: ENTV
Entertaining and informative story of how 'Toy Story 2' was almost deleted from Pixar Animation's computers during the making of the film and how the film was saved by one mom's home computer (2 mins 26 secs)
Rosenthal, D., 2014. Talk "Costs: Why Do We Care?", DSHR's Blog, Tuesday November 18 2014. Available: http://blog.dshr.org/2014/11/talk-costs-why-do-we-care.html