SDI Framework

INtune™ is a spatial data infrastructure framework that supports publishing and sharing of data in different XML schemas and different storage formats. This is done through the medium of a platform that is non-invasive or disruptive to the source or target systems, and that requires minimal changes to existing schemas or databases.

Features

• Supports the sharing of heterogeneous data between source and target systems
• Simple GUI interface for publishers and subscribers to advertise and discover sharable data
• GUI dashboard for publishers, subscribers, and administrators to view status information.
• Unobtrusive mechanism to detect data changes in a publisher’s database
• Supports ArcSDE/Oracle, ArcSDE/MS SQL Server, Oracle Spatial, MS SQL Server databases.
• Share any XML data that can be defined with an XML schema
• Supports selective sharing and selective subscriptions using open standard OCG Filter 2.0
• Supports geospatially data as specified by OCG GML 3.2
• Supports spatial filtering as specified by OGC Filter 2.0 Envelope
• Supports incremental data updates (i.e. data as they are created and modified)
• Support harvesting data with filters for precision pull of data
• Supports store and forward mechanism
• Access control policy as defined by XACML 2.0
• Audit trail of all data movements
• Supports data dissemination (one publisher to many subscribers)
• Supports data aggregation (many publishers to one subscriber)
• Peer-to-peer architecture for data exchange
• Uses JMS queue architecture that prevents data loss
• An architecture of distributed loosely coupled components connected by JMS-Q provides scalability for enterprise environments

Benefits

• No need to change existing schemas to conform to a common schema
• Does not alter the existing business processes in the organization
• Near real-time synchronization of data between publishers and subscribers
• Publisher does not have to track subscriptions for its data
• Publisher is confident that its data will be delivered to its subscribers
• Alleviates the subscriber from the need to know the physical location of the publisher
• Subscriber does not have to continuously manage requests for data.
• Subscriber is certain the data received are data that it wants
• Fine grain access control as defined by XACML 2.0
• Audit trail allows publisher to see who has subscribed to their data and the volume of data provided

About SDI Networks

An SDI network is a collection of Controller and Connector nodes that are linked together by the various artifacts needed to advertise and share data.

A Controller manages the network of Connectors and facilitates the interaction between Connectors such that they can then communicate directly and exchange information with each other. Once the source and target nodes are connected, communication becomes peer-to-peer between the Connectors, and the Controller is no longer actively involved in the exchange.

The simplest SDI network contains a single Controller that connects a data source and a data target.  Data can flow both from the source to the target, and also back from the target to the source. In more complex SDI networks, once a Controller has connected two Connectors, data can flow directly from one Connector to the other independent of the Controller.

Data Sharing

Any SDI network may be configured using any or all of the possible models for sharing data.

Managed sharing happens when data sharing between different sources and targets is orchestrated through a single node.

Aggregated sharing is where data from multiple sources is brought together in a single target.

Disseminated sharing is where data from a single source is delivered to multiple targets.

Federated sharing is where each node in a multi-node network can share data with one or more of its peers.

A harvest subscription is initiated when it is created.  It acts to “pull” data from the supplier’s database into the consumer’s database.  Harvest subscriptions are either one-time subscriptions, or can be scheduled to recur at specific intervals until a specified end time or until the subscription is retired

An event subscription is created by a data consumer.  When the data supplier makes changes in their database, the incremental updates are automatically “pushed” out to the consumer in near real time without either the supplier or the consumer taking any further action.  Changes will continue to be propagated from the supplier to the consumer until the subscription is retired.  

Data Updates

Replication and observation models describe two different ways that data targets are updated when changes are made to the data sources that they subscribe to. Replication and observation models relate to whether the history of changes is retained in the database.

An SDI network can contain both nodes that share data using the observation model, and nodes that share data using the replication model.  An SDI network can support both models within the same node using one or more destination bridges with various transformations.

Security

The SDI enforces security policies that govern the ability of members to publish and offer their data, and subscribe to the data offered by other members. Members who publish data may also impose simple security policies as part of their advertisements, restricting access to potential subscribers. Members who wish to subscribe to an advertisement must meet all security policies imposed by both the publisher and the node.

Technical Summary

Operating Environment

• Red Hat Enterprise Linux ES 4.5
• Microsoft Windows Server 2003
• Oracle 11.1
• Microsoft SQL Server 2008 SP1
• ESRI ArcGIS Server 9.3
• Java JDK 1.6
• Sun GlassFish Enterprise Server 2.1
• Apache Tomcat 5.5.28 Web Application Server
• Galdos INscape 1.8.0 (WFS 1.0)