![[Visit Client Website]](/img/gsa/banner.jpg)
SEMANTIC WEB SERVICES IN A VIRTUAL OBSERVATORY
We present a set of four web services resulting from our work on a semantic data framework in the setting of virtual observatories. These services allow a client service to search for data using three primary selections: parameter, date-time range and instrument, and also to return appropriate service links to the actual data (the fourth service). These services use a shared understanding of the inputs, outputs and preconditions as defined by a formal ontology, encoded in the World Wide Web Consortium's Ontology Web Language Recommendation and running in an internet accessible environment with Web Service Description Language (WSDL) bindings. The services can utilize reasoning services just as a user of the web portal is able to. The service client can optionally utilize the ontology when it consumes the service for additional knowledge or may be used purely syntactically (as most existing web services are now). We present these services within a specific domain context for the Virtual Solar-Terrestrial Observatory.
Introduction
We are exploring ways of technologically enabling scientific virtual observatories - distributed resources that may contain vast amounts of scientific observational data, theoretical models, and analysis programs and results from a broad range of disciplines. Simply, the main aim of a VO is to make all resources appear to be local and appear to be integrated. This is challenging because the information is collected by many research groups, using a multitude of instruments with varying instrument settings in multiple experiments with different goals, and captured in a wide range of formats. We must provide a means for an incoming user to discover, locate, retrieve and use heterogeneous and perhaps diverse or interdisciplinary data of interest. We also must provide interfaces for requests from user applications and machine generated requests for services.
We present the web services aspect of our work on semantic integration of scientific data [Fox et al. 2006] in the context of the Virtual Solar-Terrestrial Observatory (VSTO) project [McGuinness et al. 2007]. VSTO presently covers the fields of solar atmospheric physics and terrestrial middle and upper atmospheric physics. We used semantic web technologies to create declarative, machine operational encodings of the semantics of the data to facilitate interoperability and semantic integration of data. We then semantically enabled web services to find, manipulate, and present scientific data over the internet. We describe our implementation of the web service as part of our Virtual Observatory project.
Service-Oriented-Architecture (SOA)
There are notable, successful examples of enabling e-science using an internet-based service-oriented architecture using web services. The International Virtual Observatory Alliance (IVOA[1]) developed a number of simple access protocols' for application interoperability that are widel used in the astronomy community including: the Simple Spectrum Access Protocol and the Simple Time Access Protocol. Also, the Open Geospatial Consortium (OGC[2]) has, via their standards process, developed protocol standards such as the Web Coverage Service (WCS), the Web Feature Service (WFS) and the Web Map Service (WMS) together known as WxS. These services are also in wide use in applications needing integration via the geospatial coordinate system. We note these two examples in this context since they provide light-weight semantics via web services. This means that while not containing formal semantic encodings, their names and basis in either particular data-types or coordinate referencing and data product types provide hard coded semantic meaning to those clients accessing them. In essence, terms like image, spectrum, time, coverage, feature and map have a well-defined (and agreed) enough meaning in those communities to provide great utility.
Our need to provide service in an SOA environment arises from collaborations we have with the Virtual Ionosphere-Thermospere-Mesosphere Observatory (VITMO[3]) and Madrigal[4] projects.
Semantic Data Framework
An unexpected outcome of the additional knowledge representation and reasoning was that the same data query workflow is used across the two disciplines. We are finding that it seems to generalize to a variety of other datasets as well and we have seen evidence supporting this expectation in our work on other semantically-enabled data integration efforts in domains including volcanology, plate tectonics, and climate change [Fox et al. 2006b, Sinha et al. 2007].
Given the value added by this basic knowledge representation and reasoning we extended the method of access to support computer-to-computer interfaces, and in particular via the commonly adopted service oriented architecture implemented as web services.
Figure 1: VSTO architecture.
The primary query entry points are indicated at the center of Figure 1: Instrument, Parameter and Start/Stop Dates. The background ontologies, encoded in OWL-DL, are used to enable semantic filtering. At present these semantic filters constrain by physical domain (solar physics, middle-upper atmospheric physics, etc.) or an instrument or parameter sub-class (e.g. optical instruments, incoherent scatter radars, etc.). The application also uses the reasoner and input from the Metadata Service which, in this case, provides access to the date-time instances due to performance requirements there are several hundred million instances and instead we exit' the ontology and make relational database calls to a mySQL service running on a remote computer. The responses are then re-encoded into OWL for purposes of reasoning.
Figure 2: VSTO web services end-point and input example for the query interface initiating an instrument search.
The development of web service interfaces naturally followed from the web portal interface functionality Fig 1. shows the three query services: by instrument, by parameter and by start/stop date, and the Data Service which provides access to pointers to the data (in our case OPeNDAP URLs). This was the other natural service to expose as a web service and for which we have encoded the inputs, outputs and preconditions in the VSTO ontology.
We now describe the query Instrument service by example.
Figure 3: VSTO instrument query output excerpt returning OWL documents with semantic information on the available instruments according to the input selections in Fig. 2.
Fig. 2 shows the example end-point for the Query Instrument semantic web service. The web service inputs (all optional) and their types are described, and the end-point service address is given along with a link to the Web Services Description Language (WSDL ) document content for the service. Two examples are given and below that there is a Query Input form that allows a potential user of the service to scope the types of queries that may wish to make (including the semantic filters discussed earlier). Fig. 3 contains an excerpt from an example query response beginning with information about two of the 13 valid instruments.
A consumer of such a service, either another service, or client application may parse the OWL as XML without semantic meaning and use their own reasoning engine (or ours) to further work with the returned information.
Conclusions and Discussion
We currently have two clients using VSTO web services; VITMO and the Madrigal Virtual Observatory. Now that our web services are deployed at www.vsto.org, we are in a position to augment the search and query we provide in the VSTO portal by installing VSTO web services at remote locations. These services would then be accessed when a user navigates the query workflow resulting in a distributed set of queries using web services and displayed to the user.
We are also beginning work on capturing provenance. We plan to leverage the Proof Markup Language [Pinheiro daSilva, et al, 2006] an Interlingua for representing provenance, justification, and trust information. Our initial work will just use the PML-P portion of the ontology just focusing on where information came from and which services were called, later focusing on exposing the actual reasoning performed. Once captured in PML, the Inference Web toolkit [McGuinness et al, 2004] may be used to display information about why an answer was generated, where it came from, and how much the information might be believed and why.
We have also reviewed the semantic web services with respect to needs for the NSF-funded GEON project, the NASA-funded SESDI project, and the NASA-funded SKIF project and plan to add these as motivating use cases to provide the most robust and smart web service that we can.
The VSTO project is funded by the National Science Foundation, Office of Cyber Infrastructure under the SEI+II program, grant number 0431153. The National Center for Atmospheric Research is operated by the University Corporation for Atmospheric Research with substantial sponsorship from the National Science Foundation.
Deborah McGuinness, Peter Fox, Luca Cinquini, Patrick West, Jose Garcia, James L. Benedict, and Don Middleton. The Virtual Solar-Terrestrial Observatory: A Deployed Semantic Web Application Case Study for Scientific Research. In the proceedings of the Nineteenth Conference on Innovative Applications of Artificial Intelligence (IAAI-07). Vancouver, British Columbia, Canada, July 22-26, 2007, in press.
McGuinness, D. and Pinheiro da Silva, P. Explaining Answers from the Semantic Web: The Inference Web Approach. Web Semantics: Science, Services and Agents on the World Wide Web Special issue: International Semantic Web Conference 2003 - Edited by K.Sycara and J. Mylopoulous. 1(4). Fall, 2004.
Pinheiro da Silva, P., McGuinness, D., and Fikes, R. A Proof Markup Language for Semantic Web Services. Information Systems, 31(4-5), June-July 2006, pp 381-395. Prev. version, KSL Tech Report KSL-04-01.