With this Request for Information (RFI), The Open Geospatial Consortium Inc. (OGC) is soliciting input on technologies related to Contextual State Descriptions (CSD). A CSD describes the "environment" or context in which a service chain executes. The objective of this request is to identify existing capabilities that can provide a robust metadata foundation for supporting agile, adaptive, autonomous processes capable of producing repeatable, predictable results with no intervention.
In August, 2004, the Open Geospatial Consortium Inc. (OGC) issued a call for sponsors for an OGC Web Services 3 (OWS-3) interoperability initiative, a multi-threaded test bed activity to advance OGC's open framework for interoperability in the geospatial industry. Key Components of OWS-3 will Sensor Web Enablement (SWE) and Service Chaining. Each of these components are described below followed by some potential uses for CSD in OWS-3.
Organizations with relevant technology regarding CSD are urged to respond to this RFI. Information on how to respond are at the end of this RFI. The contents of the OWS-3 request for quotations (RFQ) will be based in part on the RFI response to the CSD RFI.
Sensor Web Enablement and OpenGIS SensorWeb™
A key component of OWS-3 will be Sensor Web Enablement (SWE), in which OGC members are developing standard interfaces and encodings that will enable real time integration of heterogeneous sensor webs into the information infrastructure. Developers will use these specifications in creating applications, platforms, and products involving web-connected devices such as flood gauges, air pollution monitors, stress gauges on bridges, mobile heart monitors, Webcams, and robots as well as space and airborne earth imaging devices. Several elements of SWE have already been developed and are being finalized before they are submitted to the OGC membership for adoption. CSD is a potential new element of SWE.
OWS-3 will combine the previous work on SensorWeb with the OWS-2 Image Handling and Decision support results to provide automated workflow management to produce value-added products. SensorWeb is the source of tremendous amounts of geospatial data, e.g., in-situ, imagery and simulation data. The Decision Support component of OWS-3 will continue the work of OWS-2 Service Chaining to produce value-added information products derived from SensorWeb data.
Using CSD for Service Chaining
A major thrust of OWS-2 was Image Handling for Decision Support. The Image Handling component of OWS-2 increased the ability to access and process geographic imagery into value-added products using open web services. The Decision Support component of OWS-2 demonstrated automated service chaining of imager processing services. This automation was achieved with an emerging open specification of workflow management: BPEL.
This RFI seeks input on the technology related to Contextual State Descriptions (CSD). A CSD could describe the "environment" or context in which a service chain "executes". A well-characterized service chain or workflow derives a repeatable result given equivalent sources, methods, and processing state or context. A workflow script defines the order of service invocation in a service chain but the details of the services are hidden behind the web service interface. A CSD provides information about the distributed processing environment and resources that are hidden behind the interfaces. A CSD is specifically designed to address environmental state descriptions, in such a manner that an equivalent state can be recreated at some future point to provide identical, repeatable results from a specified process, assuming identical sources, services and scripting are utilized.
When workflows are manually executed and interpreted, complexity can be addressed through human intuition and experience. However, in an autonomous environment, complexity must be distilled into codified states, before business logic can be characterized to facilitate autonomous processing with any level of confidence. Thus, a CSD is specifically designed to facilitate the autonomous processing of complex workflows in a predictable, repeatable fashion.
The promise of complex autonomous processing is the increased productivity of end user "analysts" that presently focus an inordinate amount of their time preparing data instead of analyzing it. Thus, the transition of data into information into knowledge into some form of decision or course of action is largely still a manual process that cannot be reliably repeated. True autonomous processing will not determine the ultimate course of action from a workflow, but can relieve the analyst from having to manually shepherd data through this workflow process, freeing them to focus on interpreting the derived knowledge into alternative courses of action in less time than the current process requires to develop a single course of action manually.
CSD does not describe the workflow business logic, but the contextual state that the workflow executes within. Thus, the CSD would describe the systems configuration, the hardware configuration, the firmware used by each device, the operating system, including patches, the software configuration, including external dependencies (shared libraries), and any data. All of this information characterizes the operational state or context of the system. Any change in this contextual state may result in discrepancies when the workflow is re-executed.
Potential use for CSD in OWS3
OWS-3 will apply service chaining to enable use of SensorWeb technologies for decision support. This technology will allow sensor measurements and modeling predictions to be processed into higher-value information specific to a given decision making context. CSD may be used to capture the state of the service chain including the SensorWeb components and the geo-processing services and the workflow management components. Capturing this information in a CSD will allow for post-processing analysis of the decision support information.
For instance, this initiative could describe each of the sensor components of the current SensorWeb, as well as a corresponding set of metadata to describe the processing chain. Quality metrics embedded within the data stream can describe the resultant accuracies associated with any point in the stream. However, when a degraded signal is recorded within the sensor stream, there is no context to interpret the cause and effort on the resulting value-added information. The CSD associated with this process could describe the environmental context or state that the SensorWeb was operating within to illustrate the reason for the degraded decision support information.
Due to the potential iterative nature of CSDs, a normalizing relationship structure must be provided to "relate" voluminous descriptive metadata references, without requiring these references to be redundantly embedded within a single CSD. Several compelling technologies and standards currently exist to address these specific issues. The net result of an CSD-enabled workflow would be a greater understanding of the situational context present when the SensorWeb was captured. These concepts are not new and have often been referred to as "fusion", multi-INT/multi-Source, and/or horizontal integration.
CSD could be applied to the OWS-2 Image Handling results including automated image orthorectification. In OWS-2, the orthorectification was accessible by a single web service request without exposing the underlying processing hardware. While retaining this simple web service interface, CSD could be used to capture a description of the underlying processing environment. For example, orthorectification may be performed by a cluster-based computing task with the processing load distributed across all available nodes in the cluster. Depending upon the load on the system, the number of nodes available during processing time, the system resources allocated to each node and the system, the methods and availability of source materials (raw imagery and control information), and the business logic applied to the process (tiling inputs, parallel vs. serial processing, types of control applied), the resultant orthorectified images will be created with varying degrees of accuracies and latencies. Once sufficiently populated, these accuracies and latencies (CSDs) can be mined to support the development of an adaptive system, whereby an orthorectification request can be fulfilled based on the degrees of freedom provided by varying the accuracies or latency in the fulfillment of this request.
The objective of this RFI is to identify existing capabilities that can provide a robust metadata foundation for supporting agile, adaptive, autonomous processes capable of producing repeatable, predictable results with no intervention. These objectives are purposely abstracted to describe a wide variety of processes or workflows, beyond just computer system processes.
A CSD shall describe the contextual state of a process or workflow. The structure of a CSD shall be such that a subsequent re-execution of a specified process will generate equivalent results, assuming the state from which the original process executed can be faithfully replicated. A CSD shall be capable of relating or referencing other external processes, services, or data as part of a contextual state. External references embedded within a CSD should not require the inclusion of the full reference, but rather a form of relation that will support normalized, iterative references. CSD shall not be explicitly tied to computer system processes or workflows, but must support a much broader, abstracted description of a process and/or workflow.
CSD is information, i.e., metadata, about the processing environment. OGC has adopted and implemented ISO 19115:2003, Geographic Information - Metadata in OGC Abstract Specification, Topic 11 - Metadata. While ISO 19115 is primarily for the description of datasets, it includes metadata to describe the processing history of a dataset. While the objectives of CSD exceed the elements in ISO 19115, CSD implementations must utilize elements of ISO 19115 when appropriate in lieu of defining new terms for the same concept.
Methodologies need to be developed to couple CSDs with processes and workflows as logical metadata structures. Methodologies need to be provided to extract the contextual state from a system, process or workflow and codify that state within a specified CSD. Additional methodologies need to be provided to extract a contextual state from a CSD to reconfigure (to the extent physically possible) a described system capable of re-executing the process or workflow associated with the CSD. Methods need to be developed that can identify and codify the differences between two CSDs, forming the basis for determining whether two CSDs are essentially equivalent. A CSD shall be capable of being described in a service context that will facilitate autonomous processing of CSD attributed processes and workflows.
Responding to this Request
Organizations responding to this request should send a short document to:
Executive Director, Interoperability Architecture
Open Geospatial Consortium
percivall [at] opengeospatial [dot] org
A response to this RFI must contain the following information:
- A description of the technology relevant to CSD
- Descriptions of example implementations of the relevant technology
- Applicability of the technology to OWS3, e.g., SWE and Service Chaining.
Documents in pdf format are preferred.
Submission of proprietary information is discouraged. Any response containing proprietary information must place that information in a separate file. Do not send proprietary information unless your organization has previously executed a Non-Disclosure Agreement with OGC.
Any organization with relevant technology can respond to this RFI. You need not be a member of OGC to respond. There is no reimbursement for responding to this RFI. By accepting an RFI response, OGC implies no endorsement of the technology and there is no implied agreement that the contents of an RFI response will be reflected in the OWS-3 RFQ.
For more information about CSD Services, contact:
President, Great-Circle Technologies, Inc.,
BBursey [at] Great-CircleTech [dot] com
The OGC is an international industry consortium of more than 250 companies, government agencies and universities participating in a consensus process to develop publicly available interface specifications. OpenGIS(R) Specifications support interoperable solutions that "geo-enable" the Web, wireless and location-based services, and mainstream IT. The specifications empower technology developers to make complex spatial information and services accessible and useful with all kinds of applications. Visit the OGC website at http://www.opengeospatial.org.
Candidate Submission: Tue, 2004-08-10 17:00
Close request period: Thu, 2006-08-10 17:00
TC and PC vote to issue request: Thu, 2006-08-10 17:00
Begin request period: Thu, 2006-08-10 17:00