Executive Summary
The goal of NASA's Earth Science Enterprise (ESE) is to enhance understanding of the earth system and its processes at a variety of temporal and spatial scales. At the heart of the ESE program is the Earth Observing System, a series of remote sensing instruments that will generate an unprecedented volume of data for a wide range of scientific disciplines. To manage these data, NASA is creating an experimental federation of partners. Federations provide a means for representing the interests of a broader community. As such, the federation concept is a reversal of NASA's traditional data management approach, because it places power and authority at the lowest level possible, consistent with getting the job done.
The prototype federation will consist of Earth Science Information Partners (ESIPs) drawn from academia, government, and the private sector. The ESIPs are charged with distributing and archiving scientific data and information (denoted ESIP Type 1), creating scientific products for the global change research community (denoted ESIP Type 2), and developing innovative, practical applications of earth science data for the broader community (denoted ESIP Type 3). The ESIPs Type 2 and 3 are also required, with NASA's assistance, to design and build the prototype federation over the next three years.
The latter task is difficult, because no existing federation can serve as an ideal model for an ESE federation. Federations have different objectives and structures, and consequently, different advantages and disadvantages as models for an ESE federation. To determine which lessons could be learned from existing federations, the Steering Committee for a Workshop on an Earth Science Enterprise Federation examined six federation models at a two-and-one-half-day workshop. The models were chosen to illustrate the diversity of federated structures--libraries (Association of Research Libraries and Harvard University Library), international organizations (NATO's Partnership for Peace), industry (Chevron), political (U.S. Constitution), and academic (University Corporation for Atmospheric Research). Other models, such as the Joint Oceanographic Institutions, World Data Center System, University of California System, and the Web Consortium at MIT, may also be appropriate. The models differ in their objectives, governance, potential costs and benefits, and measures of success. Based on comparison of these differences in the six federation models, and the needs of the four main ESE constituencies (data producers, global change scientists, knowledge brokers, and for-profit businesses), the Steering Committee selected the
following lessons to be considered in an ESE federation:
Lesson 1. To be successful, a federation must be a community-driven, grass-roots effort, with empowerment at the individual member level. The ESE community is broader than ESIPs Type 2 and 3; it includes many other types of data and information users and providers. Therefore, the prototype ESE federation should be planned with this broader community in mind. A step toward ensuring that the interests of all the ESE constituents are represented is to include the Type 1 ESIPs in the prototype federation.
Lesson 2. A bottom-up approach should be carried into the governance of federations to ensure that the priorities of the broader community are honored. However, some centralized management is necessary for making major decisions on behalf of the partners, for representing the federation's interests, and for conducting day-to-day operations. The instrument of centralized management, however, should be used sparingly (i.e., the ''light touch" management approach is preferred). It is essential for the ESE federation and NASA to agree on the reserve powers of the partners (that is, those prerogatives that cannot be moved to central control or to NASA).
Lesson 3. A cornerstone of federations is flexibility. In order for an ESE federation to respond to changing needs, the initial rules and procedures should not be overspecified.
Lesson 4. In an ideal federation, partners come together to achieve ends they could not achieve alone. However, since the ESIPs were chosen through a competition based on product deliverables, these common values, or the federation glue, will have to be developed by the partners. This is an essential step in forming a successful federation.
Lesson 5. It is important for any organization to decide how it will be evaluated before it is created. Quantitative metrics include measures of success and a baseline from which to measure performance on a regular basis. However, the intangible and qualitative learning that is likely to occur as the experiment proceeds is just as critical to the evaluation of the experiment. Some of the most important institutional elements are unlikely to fall within easily quantifiable categories. In the case of an ESE federation, it is incumbent on the ESIPs to determine (and NASA to agree to) the elements of this evaluation.
Lesson 6. Tensions can arise when partners in a federation have different privileges. While the ESE federation is small, equal status among prototype federation partners would help ensure that all constituents have an equal voice.
Lesson 7. There are major differences among the ESE constituents, which will lead to tensions and differing expectations. For example, there are major philosophical differences (e.g., commercialization policy) among the ESIPs. These differences must be accommodated in the mission of the ESE federation.
Ideally, a federation will increase the voice of ESE constituents, and thus their stake in the success of the program. Many of these constituents will use ESE data and information in new or non-traditional ways. A federated approach provides the flexibility and empowerment to allow them to respond rapidly to new opportunities. The payoff for greater use of the data is a greater return on the public investment in the Earth Science Enterprise.