Integration is accomplished by coupling AJAX actions with Web Services invocations and by synchronizing the actions and returned objects from the point of end users. The usage scenarios explained below use the generic integration architecture illustrated in Figure 17-C. In the usage scenarios there will be minor difference in the form of extensions. Differences come from the service specific requests created according to the service provider‟s service API (published as WSDL), or handling returned data to display on the screen. But these are all implementation differences.
Web Map Service returns maps in the form of images such as JPEG, GIF and PNG. Web Map Service clients get the maps in image formats and overlays them. Ordinary Web Map Service clients cannot use maps coming from Google Map Servers. To solve this problem and use interactive Google maps in our Web Map Service applications, we created an intermediary Google Mapping Server. This approach will also support overlaying different map layers coming from the common Web Map Service with the Google Maps ("Google Map," 2005). The intermediary server takes Web Map Service compatible requests from the Web Map Service clients, converts these requests into a new form that real Google Map Server can understand. In contrast to Open Geospatial Consortium compatible getMap requests, Google Map server uses requests with different parameters such as zoom level, tile numbers and tile width.
Evaluation of the approach: If the GIS visualization client uses Web Services from the desktop browser application and Web Services are capable of responding fast enough, then using the AJAX model for calling Web Services gives high performance increases. Since both AJAX and Web Services use XML based protocols for the request and responses, they leverage their advantages. This framework enables application developers to easily integrate AJAX based browser applications into Web Services.
AJAX and Web Services make use of XML message structures. This property allows developers to utilize their advantages together. Our proposed system enables AJAX based high performance web application approaches to utilize web services. If Web Service based applications have web based user interface for end users, then using this framework makes displays interactive. Users do not need to wait for the whole data to be received to render and display the results. Partial displaying is possible without
refreshing the whole page. Instead of making request for whole page, only the interested part will be requested. This also reduces the workload of the network traffic.
In addition to its advantages, the proposed system has a couple of disadvantages. The proposed integration framework introduces some extra work for web application developers. This extra work mostly comes from the conversion of parameters to be able to make compatible requests to remote Web Services. In order to make valid requests, the proxy server should be deployed locally, and client stubs for Web Service invocations should be created before running the application. Compared to a pure AJAX based web application, the performance of the application is reduced by the intermediary proxy server during its conversion and message handling jobs, but the gains (from our experiences) are much higher than the overhead times coming from the proposed intermediary service. This is not demonstrated in this thesis.
CHAPTER 4
FINE-GRAINED FEDERATION OF GIS WEB-
SERVICE COMPONENTS
Our federation framework provides an infrastructure for understanding and managing the production of information from distributed observation, simulation and analysis through integrated data-views in the form of multi-layered map images. Our infrastructure is based on a common data model, OGC compatible standard GIS Web- Service components and an extension of the Web Map Server, the federator service. The federator aggregates GIS services and enables unified data access/query and display over integrated data-views.
In the current context, the term “federation” means providing one global view of several data sources that are processed as one source. There are three general issues here.
The first is the data modeling (how to integrate different source schemas). The second is their querying (how to answer the queries posed on the global schema). The third is the common presentation model of data sources, i.e. mapping of common data model to a display model enabling integration and overlaying with other data sets to create an integrated data view. The first two research issues are related to lower level (database and files) data format, query, and access heterogeneities that may be summarized as semantic heterogeneity. In our research framework, Open Geographic Standards specifications for data models (GML) and online services (WMS and WFS) solve the first two issues, but we must address the third.
Our extended standard GIS Web Service components are integrated into the system through the federator service, which is actually a WMS that is extended with capability-aggregating and stateful service capabilities. These enable definition/organization of distributed data sources into shared collections, and high performance support for the responsive unified queries.
The proposed federation architecture is similar to Integrated Rule-Oriented Data System (IRODS) (Cao & Wan, 2008; Hedges, Hasan, & Blanke, 2007) in terms of research concerns. IRODS is a federated data system developed at San Diego Super Computing Center (SDSC). It is based on expertise gained through nearly a decade of applying the SRB (Doherty, Blanshard, & Manandhar, 2003; Rajasekar, Wan, & Moore, 2002) technology in support of data Grids, digital libraries, persistent archives, and real- time data systems. IRODS management policies (sets of assertions these communities make about their digital collections) are characterized in IRODS rules and state information. At the IRODS core, a rule engine interprets the rules to decide how the
system is to respond to various requests and conditions. In contrast, in our framework we don‟t need such rules, we use standard data components and their standard service interfaces defined with distributed capability metadata. Moreover, IRODS and SRB use central metadata catalog services - ICAT (or MCAT) ("MCAT," 1998) – for discovering data and services. MCAT is based on database relation tables. On the other hand, our proposed framework is based on aggregation of standard capability metadata files of the distributed data components. These metadata files can be accessed and queried remotely through the standard service interfaces. Different from the SRB and IROD, we also enable view-level federation and unified access/query over the integrated data view as a representation of shared data collections.
This section describes the implementation of view-level information presentation through federation of standard GIS Web Service components. The framework is designed for GIS domain; however we present the generalization architecture in terms of principles and requirements in Section 7.