Chapter 1. WebSphere Information Integrator Content Edition V8.2.1 architecture 13
the Figure 1-5 on page 12 scenario), the connector stages content (Step 3) to a
secure cache, and WebSphere Information Integrator Content Edition provides
an HTTP Access Servlet for the application to use to access this cached content.
1.2.2 Federation services
Federation services leverage integration services to simultaneously access
multiple sources of content and workflow automation. Federation services
Federated search for performing a single query for all relevant content across
Data maps for mapping the disparate indexing schemas of each repository.
View services for on-the-fly rendering of content.
Virtual repositories for virtually reorganizing content to support new business
Subscription event services and subscriptions for providing event notification
of changes in the repositories.
Session pools for performance and scalability
The following sections discuss each of these components.
Federated search enables a single query to search all of the content and
processes within all of an organization’s disparate sources of content and
workflow systems. A federated search leverages the search capabilities inherent
in the underlying content repositories and workflow systems.
A single query is specified and passed to the federated search system, which
then translates the search criteria into a format appropriate for each underlying
content source. It then dispatches, or brokers, the search request to each source.
The results from the search are then aggregated by the federated search system.
Federated searching may also involve post-search processing, such as sorting
the combined result set. The final result of a federated search is a single
aggregated result set that appears as if it came from a single system containing
all of the enterprise’s content.
In a federated search, a very high-level overview of the flow is as follows:
1. Access services receives the search request and determines the repositories
that will participate in the search.
14 WebSphere Information Integrator Content Edition: Planning, Configuration, and Monitoring Guide
2. Access services then dispatches query requests to each repository so the
queries can occur in parallel.
3. An instance of a server result set
is created for each search performed in the
system and is persisted for the lifetime of the client result set.
4. The server-side result set provides post-processing of the results, such as
aggregated sorting, and provides a configurable cursor-like mechanism for
result rows to be retrieved to the consuming application as they are needed,
thus decreasing network utilization and increasing performance.
Different usage scenarios will require either the federated search capability of
WebSphere Information Integrator Content Edition, or the indexed search
capability of WebSphere Information Integrator OmniFind Edition. An application
may use federated and indexed search together to provide a content integration
platform for crawling and building up indexes, and then allowing the user to take
actions in the repository on the content that resulted from an indexed search.
Data maps provide a data dictionary to map metadata across different
repositories. This data-mapping feature enables the differing metadata schemas
of each repository in the organization to be mapped to a common data model
using the data map designer tool shown in Figure 1-6 on page 15.
A server-side result set cache maintained in the EJB tier that acts as the aggregation point for
results coming back from the multiple repositories in a federated search as shown in Figure 1-3 on
Note: Unlike WebSphere Information Integrator OmniFind Edition that creates
a centralized index of the enterprise’s content and metadata that can then be
searched, WebSphere Information Integrator Content Edition provides a
mechanism for performing federated searches and working with the results
not require the creation or maintenance of a centralized full-text or
metadata index. Instead, WebSphere Information Integrator Content Edition
leverages the search capabilities of the underlying repositories to provide
real-time access to all enterprise content. Applications can then work with the
content in the search results by using the WebSphere Information Integrator
Content Edition APIs.
Note: In Figure 1-3 on page 8, federated search functionality is contained in
the “Java API” component in the Servlet Container.