ICANN|ARCHIVES
Internet Corporation for Assigned Names and Numbers
Please note:
You are viewing archival ICANN material. Links and information may be outdated or incorrect. Visit ICANN's main website for current information.
ICANN Yokohama Meeting Topic -Introduction of New Top-Level Domains: Expression of Interest #4
Posted: 9 July 2000
Letter Proposal for .geo, a New Global Top Level Domain
Dr. Yvan G. Leclerc, SRI International
6/19/2000
In response to ICANN's call for statements of interest in proposing new top level domains (TLDs), SRI International is pleased to propose the formation of a new TLD called ".geo". The primary purpose of .geo is to allow Internet users to find data by the geographic location that the data refer to. Data that refer to a geographic location or area is called georeferenced data. Examples include a 3D model of the Golden Gate Bridge, vacation photos of Yosemite taken with GPS-enabled cameras, the menu for Les Halles in Montreal, a historical document or web site on the Vatican, a weather map of the United States, traditional GIS data, or dynamic data such as the location of a cell phone or airplane.
Our proposed TLD, infrastructure, and charter will allow anybody in the world to publish, search for, and access georeferenced metadata for free and in a manner that adheres to established international standards. (Metadata contains elements such as the author, date, data type, geographic location, and URLs pointing to the data itself.) This free and open infrastructure for georeferenced metadata will enable radically new services that are not possible on the Web today by allowing users to transparently browse geospatial data by location. This capability will enable a powerful new paradigm for navigating, accessing, and visualizing georeferenced Web data.
Unlike other TLDs, the domains under .geo will not be arbitrary strings, but instead will represent geographic areas with a naming convention that clients will determine by downloading a simple XML schema from the top level .geo domain. Consequently, .geo domains will not be subject to the problems of trademark violations and the like. Indeed, the importance of users knowing domain names and URLs is drastically diminished with this architecture.
We believe that .geo warrants the creation of a new TLD because it offers a fundamentally new kind of service on the Internet, one with broad-sweeping implications and influences, one which would encompass a potentially vast distribution of data, and one which encourages a global network of users and providers of those data, much like the Web provides for non-georeferenced data today. We believe that the potential impact and openness of the proposed infrastructure deserves, and indeed requires, the recognition of a charted TLD by ICANN in order to safeguard its free and open access by contract with ICANN.
Motivation
The National Academy of Sciences estimates that 70% of web-accessible data are georeferenced. Yet, it is currently impossible for a user to find all web-accessible georeferenced data for a given area (or even a significant fraction of the data). That's because the location that the data refers to is typically not a searchable part of the data. Instead, it is either implicit or part of associated metadata. Thus, today's search engine technology is not applicable.
A number of companies and organizations have created private or governmental databases that hold a small fraction of all georeferenced data. Companies like Mapquest maintain private map databases with associated street addresses and pointers to businesses like restaurants and shops. Organizations like the Federal Geographic Data Committee maintain government-owned imagery and coordinate data databases that can either be accessed directly or via a Clearinghouse that let's a user find georeferenced data across many databases through a keyword-based search engine. The FGDC Clearinghouse is an example of a database that uses metadata containing the geographic location and URLs of the data.
None of these organizations is either capable of, or willing to, create a database that would allow anybody in the world to publish and search georeferenced metadata for free. Indeed, this task is so large that no single, regional, organization could do it alone.
What is needed is a coordinated global infrastructure with participating organizations from around the world, as we propose below. The proposed .geo chartered TLD will greatly enhance the probability that such an infrastructure can be built and sustained globally. It would also enable existing search engine and mapping companies to search for georeferenced metadata with minor modifications to their software architectures. New markets for viewing and analysis of this data would emerge and add value to existing commercial web capabilities.
Technical Approach
We propose to build and maintain an open standards-based infrastructure on .geo that will enable anybody to publish and search for all metadata referring to a given area for free. The infrastructure is based on a hierarchy of servers whose domain names represent geographic areas, nominally of the form minutes.degrees.tendegrees.geo. The exact form of the naming convention will be available as a simple downloadable XML schema from the top level .geo domain. No other names are anticipated beyond a few administrative domains.
For example, consider a server with DNS name 10e20n.geo. Since it is a server at the second level of the hierarchy (tendegrees.geo), it is responsible for a 10 degree x 10 degree "cell" of the world. The service area of the cell spans from longitude 10 degrees East and latitude 20 degrees North to longitude 20 degrees East and latitude 30 degrees North. Similarly, server 1e5n.10e20n.geo is responsible for a 1 degree x 1 degree area. Metadata is placed at the appropriate level and position in the hierarchy as a function of the geographic footprint and location of the corresponding data. See http://www.dgeo.org/ for a prototype DNS hierarchy and server that we have built as part of our DARPA-sponsored Digital Earth project. This hierarchy and related software will serve as the basis for our proposed .geo infrastructure.
This naming convention and metadata placement strategy allows clients to determine which host(s) to query for metadata referring to a given area, thereby distributing the load over many servers with no single point of failure or congestion. This is a significant advantage over current approaches (such as the FGDC Clearinghouse) in which all client queries pass through a single host. An open source API for publishing, querying, and accessing the metadata will be made freely available.
The use of DNS also allows metadata to be transparently moved to new hosts and subdomains as needed. For example, initially all metadata may be physically hosted on a single computer, with all .geo domain names aliased to that machine. But over time, as areas fill with metadata (such as large urban areas), their corresponding subdomain and metadata can be transparently transferred to new computers. These computers may be maintained by other organizations who may, when necessary, transfer data to new computers (and new organizations) deeper in the hierarchy.
This distribution and query method drastically reduces the storage and bandwidth requirements of any given server, making it possible for thousands of high-speed visualization systems to be simultaneously searching for and retrieving metadata and data.
The .geo hierarchy and distributed infrastructure enables a radically new approach to finding, visualizing, and interacting with data on the web. Instead of searching for data in Web sites or entering keywords into a search engine and looking through the hundreds of resulting links, a user can specify a profile and then simply navigate in 2D or 3D and have the visible data presented within the spatial context of a realistic model of the earth (which will itself be represented in the .geo hierarchy). Spatial data freely available today can, and has been, used to provide a very adequate baseline to build a world model to enable this new browsing paradigm. Of course, a standard textual search engine augmented with a location criterion is still possible.
Many other uses for this free and open worldwide index of georeferenced data can be imagined. For example, a complete index of the geographic location of IP-addressed devices, or static and mobile phones, becomes not only possible but quickly accessible.
Business Model
The .geo hierarchy, were it to be expanded to the smallest cells, would require tens of millions of servers and terrabytes of storage. It would also require appropriate supervision to ensure that the free basic services (publishing, searching, and disseminating of any metadata) are being offered compliant with the appropriate standards and performed adequately.
Such an enterprise cannot be sustained without significant financial resources, which are clearly not available from the free basic services. We propose that such resources be obtained through optional, non-free services to data providers. Such optional services would include caching and/or hosting data, and providing various security mechanisms for restricting access to the metadata and/or the cached or hosted data.
These optional services would be provided by the franchise owners of the particular domains, who would pass on a fraction of the data-provider fees to the parent domains (ultimately .geo) whose responsibility is to maintain the integrity of the hierarchy beneath them. Domain ownership would be allocated on a franchise basis to qualified commercial or government organizations.
Charter
The charter of the founding organization for .geo, which we have tentatively called the GeoNIC, would be to define and enforce standards and practices for hosts operating within the .geo hierarchy, including:
- registering domains (cells) in the hierarchy to qualified commercial or government organizations;
- enforcing the requirement that cell owners provide free access to the basic services (publishing, querying, and accessing metadata) for all users and data providers;
- enforcing adherence by cell owners to appropriate international standards (e.g. ISO TC 211) for metadata, data, and access to services;
- participating in the international bodies that define such standards;
- providing a framework for testing conformance and ensuring that performance targets are being met; and
- generally ensuring the integrity of the hierarchy.
SRI International as Founding Organization
SRI International is a non-profit research institute that is uniquely suited to be the founding organization for the proposed top level .geo domain. Its 2,500 researchers, scientists, technologists and business consultants recently celebrated 50 years of scientific discovery, new technology applications and innovative management approaches that have benefited businesses, governments and organizations worldwide. Its original charter -- to "promote and foster the application of science in the development of commerce, trade and industry [for the] prosperity of mankind" -- remains its mission.
As important as its charter and wide-ranging research, engineering, and business experience, SRI has deep roots in the Internet:
Arpanet
SRI was one of the first two nodes of the Arpanet.
NIC
SRI served as the Network Information Center (NIC) for over two decades.
Furthermore, our team has wide-ranging experience in the creation and use of standards for georeferenced data:
Digital Earth
SRI's Digital Earth project is currently being funded by DARPA to create a testbed hierarchy that will serve as the basis for the worldwide .geo hierarchy.
TerraVision
TerraVision is our freely available, multi-platform, network-based terrain browser. It is currently being re-engineered to use the .geo hierarchy.
GeoVRML
SRI was the founder of, and is the primary contributor to, the GeoVRML Working Group of the Web3D Consortium. GeoVRML is an extension of the ISO standard Virtual Reality Modeling Language (VRML97) that enables the creation of georeferenced 3D models. GeoVRML 1.0 has recently been accepted as a recommended practice by the Web3D Consortium and is being considered for adoption as an ISO standard.
Participation in Standards Organizations
SRI is an active member of the OpenGIS Consortium and the US Federal Government Digital Earth Working Group, and is formally collaborating with the Alexandria Digital Library Project.
When the size of the .geo hierarchy grows beyond certain limits, SRI will transfer management of this hierarchy to a dedicated subsidiary. However, this would only be done with ICANN's approval and guarantees on adherence to the founding principles of open standards and free access defined in the charter.
Proposal Team
Our team consists of technical and management staff of SRI International, whose headquarters are located at 333 Ravenswood Avenue, Menlo Park, CA 94025. David Scott is located at the Washington, DC, office, located at 1611 North Kent Street, Arlington, VA 22209-2111.
The point of contact for technical issues regarding this proposal is Yvan Leclerc. The point of contact for business issues is Michael Summers. Please do not hesitate to contact either of us (or any other member of the team if we are not available) if you have any questions about this proposal.
The team currently consists of the following individuals.
Dr. Yvan G. Leclerc
Sr. Computer Scientist, Artificial Intelligence Center
Principle Investigator, Digital Earth Project
TerraVision Designer
Mr. Michael Summers
Vice-President, Business Development, Software
Dr. Martin Reddy
Computer Scientist, Artificial Intelligence Center
Investigator for SRI'S Digital Earth Project
Co-chair of the GeoVRML Working Group
TerraVision Software Engineer
3D Visualization Expert
Dr. Lee Iverson
Computer Scientist, Artificial Intelligence Center
Investigator for SRI'S Digital Earth Project
Co-chair of the GeoVRML Working Group
TerraVision Software Engineer
Open Source/Standards Expert
Dr. Michael Eriksen
Computer Scientist, Artificial Intelligence Center
Investigator for SRI'S Digital Earth Project
Digital Earth Infrastructure Software Engineer
Mr. David J. Scott
Program Manager
GIS Expert