Developing an LTER Information System for the 21st Century
DRAFT version Jan. 10, 1996, Input from IS and DataTask groups
pending
Table of Contents
Date: Wed, 10 Jan 1996 11:24:28 -0700 (MST)
From: "James W. Brunt" <jbrunt@sevilleta.unm.edu>
To: is@lternet.edu
Cc: datatask@lternet.edu, datatask@lternet.edu
Subject: NIS plan for Network Office Proposal
Good Day Folks, -
I've taken what John wrote up for the CC meeting, what Rudolf has put together, what I wrote for the last proposal, and what Jim has written, and put it in a pot and tried to distill something that I felt would survive scrutiny. Well, I'm quite sure I've stuck my neck out here but hen I guess that`s what necks are for. Attached is a first draft of the NIS development section for the network office proposal. I've embelished and chopped to begin to mold it into something that I think is fundable.
The catch: I've left a lot of material in this text as explanatory and overly descriptive for your review and for the executive committee review this weekend. Most of it is going to have to go. I've been told that there might be room for 2 pages of text in the proposal. However, we can still put as much information up on the web as we can produce.
Therefore, I suggest that you try to turn this document into one that you can live with and that can be posted on the LTERnet IM web page as a full and evolving description of our NIS development. The document that is currently on the web is simplistic in our LTER kind of way and I don't feel it suggests to the reader what we and the community already understand about information system development. It's got a lot of language in it that tiptoes around technophobe researchers. I don't think that language will survive review today. It suggests of an LTER that existed 10 years ago. So, I've tried to modernize the approaches in places. I'm prepared to be cutdown but I'm not prepared to sacrifice the success of the proposal by being overly cautious.
It needs a lot of work so I hope you will all take time to scrutinize it carefully. Please post your responses to is@lternet.edu.
Rudolf, please post this on the LTERnet server and let everyone know where it is.
James
James W. Brunt
Sevilleta LTER Project
Department of Biology E-mail: jbrunt@sevilleta.unm.edu
University of New Mexico FAX: (505) 277-5355
Albuquerque, New Mexico 87131 Phone: (505) 277-9342
----------------------------------------------------------------------
*B. Infrastructure Development (back to TOC)
*Network Information System
Research Data Management at U.S. LTER sites is carried out with common objectives in mind, but under a variety of different circumstances and constraints. There is a great deal of heterogeneity among sites in the approaches taken to meet research objectives. Meeting standardized goals with a variety of solutions has built strength into this system. Data Management protocols within LTER are built on previous investigator experiences in the International Biosphere Program (IBP) and have capitalized on the easy availability of computing power. Although, computers have come a long way, the fundamentals of managing research information and developing an appropriate system for management have remained relatively stable. What has changed drastically in the last decade is the expectations of the researchers and the community. With the explosion of internet connectivity and the birth of the World Wide Web (WWW), scientists, administrators, and the general public have come to expect greater access to the products of valuable research dollars.
*Introduction to Current System (back to TOC)
Since the first publication of LTER Catalog of Core Data Sets in 1990 scientists within the LTER network as well in the wider ecological research community have been aware of the wealth of long-term data that are being collected by LTER investigators. During the same period of time that the LTER data managers developed the Catalog, the work of the LTER Connectivity Committee and subsequent efforts by the sites to implement its recommendations resulted in near full Internet connectivity at main administrative locations. This created the foundation of an infrastructure for electronic communication and data exchange that would substantially enhance the traditional means of networking through meetings, conferences, symposia, etc. Based on this rapidly evolving infrastructure, electronic versions of the Catalog and the of the LTER Personnel Directory became available to most researchers in 1991. The evolution of those data repositories from hardcopy versions to on-line databases accessible by e-mail, FTP, Gopher, WAIS, SQL and Web information servers reflects the successful use of the latest communications technology in the integration of individual sites into a coherent research network. When the LTER All-Site Bibliography was created in 1993, a hardcopy version was no longer considered an economical or even desirable implementation. Instead, it was implemented as a searchable online database from the very beginning.
Improved accessibility and ease of use highlighted existing gaps in the LTER information infrastructure. While the Gopher and Web information servers installed at LTERnet in 1992/93, together with servers installed subsequently at most sites, simplified access to information that is widely geographically dispersed, queries are still unwieldy. In many cases (other than with Core Dataset Catalog, Bibliography and Personnel Directory), to find out what data is available on LTER site information servers, you need to make individual connections to each of the site servers.
Many LTER datasets are available via WWW and other network information servers, but to date it is difficult to analyze and synthesize data from different sites because most site servers return data in different formats. This situation prompted the LTER data managers at their 1994 Annual Workshop to commit to building a system that facilitates cross-site data exchange for intersite research, the LTER Information System for the 21st Century.
The LTER Network Office will foster expansion of
the existing LTER-wide Information System. This will include substantive
participation in the development of advanced query and information
systems which integrate data from the individual site information
systems. To further this end, it will support planning activities,
aid in coordination of site activities, promote standards development,
and provide access to software, storage and network resources.
*Philosophy and Goals for Development of an LTER-Wide IS (back to TOC)
The primary function of the Network Information System is to support intersite research. All technical considerations will be driven by this goal. Recognizing that a basic information system has been in place and available to LTER researchers for some time (with components such as Catalog, Personnel Dir., Bibliography, integrated mail forwarding and direct links to most site systems), it is imperative that any future LTER Information System will be developed while maintaining and expanding on the existing functionality. A modular, step-by-step approach will ensure that existing functions will be seamlessly integrated with the future system. Information management systems are always evolving, due both to technological changes and to improvements in our understanding of how scientists most efficiently use data. Recent revolutions in computer and network technologies are reshaping and expanding the types of feasible information management systems.
Goal 1: Increase Utility of Existing System
With the objective of fulfilling its mission in information management, the Network Office will be responsible for further development of the LTERnet Network Support System to maintain andexpand the following core activities:
Goal 2: Increase Access
Goal 3: Maintain Local Autonomy
Develop from a Research Perspective
Because development of the system must be driven
by the needs of ecological scientists, it is logical to coordinate
the development closely with efforts by various LTER intersite
and synthesis research groups. Each group leading an LTER intersite
research effort will include at least one member from the LTER
information management group. This will help the research groups
to stay closely informed on the present and planned capabilities
of the Information System and provide an effective channel to
communicate their needs and expectations to the information managers.
Initially, intersite research groups will provide the main pool
of users of new system functions and they will provide the feedback
required for testing.
*System Characteristics (back to TOC)
The Network Information System working group has developed the following functions for the LTER-wide IS based on input from the data management committee and researchers in attendance at the Snowbird meeting in 1995. An LTER-wide IS should facilitate:
In addition, the system should integrate site information systems, not replace them. The envisioned system will be a distributed system using advanced network tools that will ensure the system is useful into the coming decade. For example, network client/server technology is well suited to this application and can be implemented independent of the specific computer platform. Some de facto standards that can help ensure hardware and software independent functionality have emerged in this area and would be considered in our design (e.g. ODBC). There are also some recent developments with potentially major implications for the way future network information systems are implemented independently of any specific hardware or software platform. In particular, executable content on Web servers by the Java language (an open standard for programming on the Web) could be used to package data together with their metadata.
*Plan for Development of an LTER-Wide IS (back to TOC)
Design and implementation of the IS in a modular, step-by-step fashion will require that teams of information specialists accomplish specific system parts in a coordinated way and in a predictable period of time. Such task will include implementation of access and retrieval functions for datasets that are standardized in terms of content and apparent format. For example, measurement procedures of net primary productivity can vary considerably among such different ecosystems as lakes (NTL) and Forests (AND, LUQ) and deserts (SEV, JRN). Correspondingly the data reflecting those measurements may not be comparable. In this case, a team of information specialists would work with the sites to add to the IS the capability to access from all sites productivity data and metadata in a consistent format, with consistent tools.
The development of a distributed system will depend upon the readiness of the site to participate but will not exclude the site's information. Sites that do not yet have full capabilities can have their contributions supported by the Office server.
The development of an LTER-wide Information system will depend on the efforts of many critical players: The LTER Coordinating Committee, site researchers, Network Office Staff, Data/Information Management Committee, and Collaborators (e.g., SDSC). These efforts will be focused at a continuing series of workshops that will form the basis for the design, funding, and implementation of an LTER-Wide IS.
Annual meetings of the Data Management Committee and the NIS working group will serve to coordinate the effort. NIS task groups composed of researchers, data managers, and occasional consultants will provide the labor necessary to design, fund, and implement various components of the system.
The NIS working group will solicit interest from a couple of intersite research groups and hold workshops to assess researchers needs - what works, what doesn't. These workshops can be organized in conjunction
with workshops that investigate specific intersite research questions, such as climate patterns, biodiversity distribution, litter decomposition processes (LIDET) or Hydrology. The CC will determine which areas to cover first and the subjects initially chosen would relate to core areas. The workshops will also address and attempt to resolve outstanding standardization issues that are relevant to how data and information is presented and distributed by the Information System.
Data of interest to the group, such as the Climate Committee will be used by a task group to demonstrate present capabilities. For example, compile a Web page of all climate data available on LTER site servers, download them into spreadsheets or other applications, and graph them. Then, the NIS task
The group will then compile a list of problems and
deficiencies with the present approach. Why is it insufficient
in supporting the groups intersite research? Propose solutions
to fix the shortcomings. Develop an information model for the
data. Set time table for implementation. The process would then
proceed with mechanisms built in for testing and feedback.
Workshops:
Goals for the year:
Workshops:
Goals for the year:
Workshops:
Goals for the year:
Workshops:
Goals for the year:
Workshops:
Goals for the year:
Workshops:
Goals for the year:
Workshops:
Goals for the year:
Resources Needed from LTER Network Office (back to TOC)
The Office will foster expansion of the existing LTER-wide Information System. This will include substantive participation in the development of advanced query and information systems which integrate data from the individual site information systems. The Office funds annual meetings of the Data Managers Working Group where continued development will occur on this important issue.
As introduced under Synthesis, the Office will fund
a lead person for each of the Network Data Sets to work with LTER
sites to develop standard formats for archival and access to data
in the site information management systems.