and store digital photos through Flickr, employ Apple’s Time Capsule for regular personal computer backup, and use LexisNexis for online legal services.

The commercialization of data storage and services contributes an important component of the data CI environment needed to harness the potential of our information-rich world. However, private-sector storage and services are not the solution to all digital data needs. For some digital data considered to be “in the public interest” (such as census data, official records, critical scientific data collections, and a variety of irreplaceable data), a greater level of trust, monitoring, replication, and accountability is required to minimize the likelihood of loss or damage and ensure the data will be there for a very long time. For such community data sets, stewardship by a trusted entity (such as libraries, archives, museums, universities, and institutional repositories), whose mission is the public good rather than profit, is generally required.

There is no one-size-fits-all solution for data stewardship and preservation. The “free rider” solution of “Let someone else do it”—whether that someone else is the government, a library, a museum, an archive, Google, Microsoft, the data creator, or the data user—is unrealistic and pushes responsibility to a single company, institution, or sector when what is needed are cross-sector economic partnerships. Sustainable economic models for digital data in the public interest are the focus of an international Blue Ribbon Task Force for Sustainable Digital Preservation and Access ( brtf.sdsc. edu), whose goal is to examine digital preservation as an economic activity and explore cost frameworks for various institutional scenarios. The Task Force’s final report, due at the end of 2009, will focus on economic models, components, and actionable recommendations for sustainable digital preservation and access, though it is already clear is that a diverse set of economic approaches are necessary.

In aggregate, these four trends point to the need to take a comprehensive and coordinated approach to data CI and treat the problem of sustainability holistically, creating strategies that make sense from a technical,

We do not produce
storage capacity
at the same rate
we produce digital
information. even
if we wanted to,
we cannot keep all
of our digital data.

policy, regulatory, economic, security, and community perspective.

value and sustainability

In developing effective models for data CI, perhaps the greatest challenge is economic sustainability. A key question is: Who is responsible for supporting the preservation of valued digital data? Critical to answering is the recognition that “value” means different things to different people. There is general agreement that official digital government records (such as presidential email and videos of congressional hearings in the U.S.) are preservation-worthy and of great political and historical value to society, but the video of your niece’s voice recital is likely to be of value to a much smaller family group (unless, of course, your niece is, say, Tina Turner).

Sustainability solutions for digital data are inextricably related to who values it and who is willing to support its preservation. Governments worldwide are willing to support the preservation of digital content of national value, a substantial undertaking that involves hosting multiple copies of the same data, migration of the data from one generation of storage media to the next to ensure it lives in perpetuity, and protection of its integrity and authenticity. Your niece’s voice recital may live on the hard drives of one or more family members, but there is rarely an explicit plan for how such a treasured family artifact will be preserved for the next decade and beyond.

How might we distinguish among all the data use, stewardship, and preservation scenarios to create and identify the data CI solutions needed to support them? One way is to borrow from the world of computation and adapt the Branscomb Pyramid model to today’s data-use and data-steward-ship scenarios. In 1993, the NSF asked Lewis Branscomb to lead a distinguished committee to consider the future of high-performance computing in the U.S. The final report included a useful framework, now known as the Branscomb Pyramid, where the base of the Pyramid associated the least-pow-erful computational platforms with users needing computation for “ everyday” applications, the middle associated more powerful computational plat-