Interoperability Makes Sharing Data Easier than Ever Before

The tantalising possibility of letting a variety of users unearth and share useful information in real time has piqued the interest of many organisations in modern business intelligence applications. But incompatible IT systems, highly specialised software and inaccessible data can hinder or sabotage this potential, as well as sink the value of previous technology investments.

That's why interoperability - a broad concept that encompasses the capability of different software and hardware from different vendors to share data and cooperate in dynamic processes - has become so important in IT environments and why it is a central attribute of SASŪ9.

The interoperability of SAS software is apparent in a multitude of areas, at both the data level and at the dynamic-process level. At the data level, this includes shared data stores from a variety of sources; SQL, MDX, ODBC and JDBC query support; predictive model sharing; and metadata. And at the level of dynamic processes, interoperability enables message passing, distributed-object interfaces, Web services and publishing via the enterprise event bus.

Using Open Standards for Data Interoperability

SAS continues to expand the types of files it interoperates with, says Steve Rees, Senior Architect at SAS UK. For instance, SAS' new XML interchange engine enables the rapid import and export of XML documents. SAS can also import and export files in formats such as Microsoft Excel, CSV, tab delimited, custom delimited and Lotus.

For users of many popular clients, data interoperability enables access to SAS cubes and tables. With the new SAS OLAP Server in SASŪ9, Microsoft Windows-based clients can access SAS OLAP structures because SAS supports OLE DB for OLAP, an MDX query language. Drivers for SQL query languages JDBC and ODBC allow clients to use SAS data tables.

Accessing data in market-leading relational databases and enterprise resource planning systems is simple in SASŪ9. RDBMS adapters for Oracle, DB2, Microsoft SQL Server and Microsoft Access are available, as are adapters that provide access to ERP systems such as R/3, Siebel and PeopleSoft.

"Shared data sources, access engine infrastructure and the ability to read different file formats are not really new to SAS. They are part of SAS' longstanding interoperability," says Rees.

Interoperability also extends to predictive models in SASŪ9, which means that users can share models and deploy them in environments other than SAS, says Maria Charlesworth, Technical Product Manager Analytic Intelligence.

The medium for this interoperability is Predictive Modeling Markup Language (PMML), a standard developed by the Data Mining Group, an independent organisation of database and data mining vendors in which SAS is a leader.

SASŪ Enterprise MinerTM can store its data models in PMML These models can then be exported to a database that also supports PMML, such as IBM's DB2, and can be scored within that database - without having to move data around.

Extending Interoperability to Metadata

Metadata, or information about data, has become increasingly useful, particularly in administering information technology - for seamlessly managing servers, users and security, for instance.

The SAS Metadata Server, which centrally stores and secures metadata, helps manage and track data for data warehousing. "What the metadata layer provides is much more structure, rationality and visibility in building and exploiting the data warehousing environment," Rees says.

The sharing of metadata among applications may be a major hurdle for organisations, however. To address this interoperability challenge, SAS has adopted the Common Warehouse Metamodel standard.

This makes it easy, for instance, to import existing metadata into the SAS platform. "There is no need to redefine a warehouse environment that has already been defined. Setup time and administration is reduced as you don't need to rebuild the metadata," says Rees.

Enabling Web Services with Dynamic Interoperability

One of the big payoffs of SAS' interoperability is flexibility in the ways users can talk to the SAS back-end.

"Callers to the SAS platform can use SAS from a variety of clients. The trend is toward this looser coupling of the caller with the back-end," says Dave Annis, Head of Technology at SAS UK. "What it means is that more users can take advantage of SAS' 900 or so functions and approximately 300 procedures."

For an application to dynamically use services provided by the SAS back-end, what's necessary is a form of interoperability that enables a request to be sent and a response to be returned, Rees explains. SAS implements several standards that support this process, ranging from distributed-object interfaces, message-oriented middleware and the new standout in the area of loosely coupled service-oriented architectures: Web services.

SAS supports Web services through Microsoft's .NET Framework and the Java API for XML. Using Web services, organisations can provide authorised users access to SAS analytics regardless of their locations or their own IT setup.

"Web services is a mechanism that is still evolving, but it's another interoperability standard for SAS," says Annis, noting that SAS participates in the Web Services Interoperability standards organisation.

Efficiently Sharing Crucial Information with Event Publishing

Interoperability at the dynamic-process level is also apparent in the new event publishing capability in SASŪ9. The enterprise event bus can receive alerts and notices and broadcast in other applications immediately.

For example, if analysis by a forecasting application discovered that inventory was being depleted more rapidly than expected, this alert could be published via the event bus and more inventory could be ordered right away, says Rees.

"It used to be that you'd have to forecast, print the report and then share the information. Now you can provide that key information in real time," Rees says.

For more information about integrated technologies, including samples, papers and downloads, visit the Enterprise Integration Community.