Architecture for Interoperability
ARCHITECTURE FOR INTEROPERABILITY
A project funded by the U.K. Ministry of Defence is seeking to understand
how consistent and unambiguous geospatial data can be provided
from geospatial providers to the exploiters of the information.
by Steve Little
A project funded by the U.K. Ministry of Defence (MOD) is seeking to understand how consistent and unambiguous geospatial data can be provided from geospatial providers to the exploiters of the information, end users or systems, and how international interoperability can be achieved. The initiative is being carried out by Envitia (formerly TENET Technology) under a three-year research contract awarded in 2007, the Geospatial Intelligence Integrated Reference Architecture (GI2RA), to investigate the future provision and dissemination of environmental or geospatial information and intelligence.
Envitia is building on previous work undertaken in providing research and advice into how a service-oriented architecture (SOA) can be used to support the provision of environmental or geospatial data. This included data modeling and data conversion of geospatial data from a number of different nations into a product-neutral and standardized form, to support interoperability and portability.
The provision of geospatial information and intelligence from geospatial providers to end users or systems has been predicated and funded for specific user or system requirements. Consequently the content and format of the data (the products) are tailored towards these users or systems. Where exploiters of the information want to consume information provided in different products or from a number of providers, at a national or international level, a number of issues need to be addressed.
The more tangible issues include the fact that there are a large number of products in a number of different data formats across all domains. To support joint or multi-domain operations, systems or users typically require access to different products, in different formats. There is an overhead on geospatial providers and exploiters to support these different products or formats, and understand the differences between the different formats. Another challenge concerns inconsistent definitions. Users or system integrators not only have to understand the different formats and/ or products, and their limitations, but also must understand the content of the products. As different products have evolved within different domains, the definitions for particular objects have evolved.
In addition, there is the issue of duplication of data. Within different products, which single systems or users might need to ingest data from, a number of real-world objects or features are repeated, often with contradictory attribution or values. If the user is presented with two coastlines, one from an ENC, one from a DNC, which one should he use? It would be better if the user did not have to make the choice, and only given one, appropriate to how it is to be used.
Within individual nations and internationally across coalition forces where a considerable amount of data is held, there is also an issue over how users or systems who require access to the data know what data is available. It was suggested that for a recent operation, coalition forces between them held 95 percent of the available geospatial data. But no single unit or organization was aware of the data held by other units or organizations. A considerable amount of time and resource was expended on collecting geospatial data already held within the coalition forces.
RECOGNIZED PICTURE
The issues discussed above are not new, and a number of nations have undertaken research and programs to support the future provision of consistent and coherent geospatial information and intelligence. The approach currently being undertaken by the British has evolved over the last 10 years. The MOD has sought to provide a recognized environmental picture (REP) in support of the joint operations picture, defined as being “a controlled database comprising dynamic and quasi-static geospatial and environmental information, referenced to a common geodetic framework and tailored to current operations. It will provide all users with access to consistent maps, charts, meteorological forecasts, aeronautical information and remotely sensed data represented in a seamless digital representation of the joint battlespace.”
The provision of an organization to manage a centralized REP database was discounted, due to organizational and financial constraints. British officials decided instead to undertake research, through GI2RA, for the development of an open IT architecture and principles to coordinate production, validation and management of information from geospatial providers with their counterparts to ensure that consistent and fused geospatial information is provided.
The research to be undertaken as part of GI2RA is split into three phases. The scoping study provided the scope and boundary for the research, supplying a clear focus to the research activities, particularly with respect to the first year. The scope of GI2RA, in relation to other MOD (including geospatial providers) research projects, coalition partner activities such as the U.S. Department of Defense, and evolving open standards and technologies, was considered. The operational impact of not adopting the research was also considered.
The databases and exchange work package is undertaking research to determine an extensible logical and physical model to support the future provision of environmental information. The research is investigating possible architectures and understanding how the information can be updated or exchanged. The information model being developed will document the terms, abstractions and entities required by users or systems to exploit the data. To support interoperability and to bring some sort of standardization to the provision of this data, the ISO 19100 set of standards is critical to the research.
Finally, the information bearers work package has two main threads. The first is the development of compression techniques, as well as information management, information assurance and information exchange techniques, while the second will address how to automatically replicate the information across different security domains. The research must determine how intelligent use of the available bandwidth and the existing or planned Infrastructures can be made.
TRIALS/TESTBED
In addition to the three distinct phases of the research, a persistent trial environment is being established. The testbed will be populated with relevant data and used to look at some of the issues around the practical fusion of geospatial data. The trial will set up and test approaches to compression, replication and security using standard Web-based technologies.
GI2RA was demonstrated at the Coalition Warrior Interoperability Demonstration (CWID) in 2008, where a limited amount of instance data was fused. GI2RA will be demonstrated at CWID 2009, where it will be more integrated into the provision of geospatial data. The approach being undertaken for the information modeling is to produce a set of geospatial models that are complete, coherent, accurate, useable and extensible. We are adopting the ISO approach, which is to segment the description of elements required to make a geospatial reference model to support instance data. They are feature concept dictionaries, feature catalogs, conceptual application schemas, exchange schemas and coordinate reference system dictionaries and units of measure dictionaries, and likely data type dictionaries.
The goal of the information modeling process is to ensure that all these elements are consistent for GI2RA, and consequently for all users of geospatial information through the GI2RA model.
Within the source data model features, attributes and enumerations are modeled in their original, product-specific form, encapsulated in an open-standards carrier format rather than a set of proprietary or legacy carriers. The source coordinate reference system, time domain and vertical reference frame will be preserved. These features and attributes can be accessed from the source model through Web services, converted to the required formats.
Information from the source data model will be modeled as a harmonized, product-neutral form. Rules are being developed to relate source and harmonized components so that data can be brought into a harmonized view. To harmonize the data, the model needs to support a one-to-one mapping of features and attributes and also more complex mappings, such as splitting a feature type based on the value of an attribute into many feature types, or fusing feature types together and forming an attribute that captures their different categorization.
The information held in the harmonized data model will be transformed to a common coordinate reference system, time frame and vertical reference frame, providing a consistent, coherent, unambiguous view of the geospatial information to the exploiter.
The rules to support the harmonization of the source data model could be independent of the particular instance data. This means that the data sources to be harmonized could include individual nations’ geospatial providers or data from a number of coalition partners.
PRACTICAL REALIZATION
Geospatial providers typically have their own silos of geospatial information. These will continue to be maintained within the existing organizations or nations. But the elements required for the coherent and consistent geospatial reference model will be provided through registry services, as described by the Open Geospatial Consortium. An approach being considered is for “on the fly” harmonization of data from the existing silos, at either a national or international level as appropriate.
At both the national and international level, the harmonized model will use consistent definitions for the information it describes. The information can be transformed into a common format to support what is required of the data. Since duplicate instance data has been harmonized, the user will be given a consistent view of previously repeated instances. The registry services will enable the geospatial information to be efficiently searched and discovered.
The GI2RA research might determine that current information systems cannot support on-the-fly harmonization of geospatial data to support the data fill for deployed users. However, the approach being taken—of modeling different sources of geospatial information and harmonizing these different sources—could support not only national but also international interoperability. ♦
