Business intelligence tools have become mission-critical and, as a result, there is growing demand for wider accessibility across the organization.
Business intelligence tools have become mission-critical and, as a result, there is growing demand for wider accessibility across the organization.
In fact, business users want BI to be part of the world in which they work – to integrate with their company portals and enterprise search capabilities and be accessible via mobile devices and embedded in process flows and other applications.application for individual groups. It must fit into existing, varied and complex infrastructures.
That means one thing: BI can no longer be implemented as a stand-alone
But that is difficult to manage effectively given multiple applications, platforms and data sources must be taken into account when implementing BI. The best bet is to use a service-oriented architecture (SOA) to ease integration.
SOA is an approach to software design based on seven key principles that map nicely to the needs for efficient, reliable, agile enterprise BI solutions. These principles help unlock the value behind enterprisewide BI deployments.
* An open, standards-based architecture
First, SOA is open and standards-based. Web standards such as Simple Object Access Protocol (SOAP), Web Service Description Language (WSDL), and Web Services for Remote Portlets (WSRP) are key to achieving integration.
The SOA standards provide organizations using an SOA-based BI solution the flexibility needed to leverage their existing IT infrastructure without having to duplicate functions, such as security, or deal with issues associated with proprietary application servers. An SOA-based BI solution can run on any Web server and complies with existing routers and firewalls. A single application program interface (API) ensures the solution integrates easily with other programs and can be extended to meet business needs that transcend BI.
* Platform neutral and location transparent
SOA is platform neutral and can leverage any existing infrastructure. Organizations can run SOA-based software on existing hardware, databases and application servers.
SOA also is location transparent, meaning services can be located anywhere on the network. This type of flexibility is important considering the nature of BI means reaching data and users dispersed across networks and geographies. For example, IT might want to locate query services close to large data centers to avoid network latency when processing incoming requests. With location-transparent services, IT adopts the most efficient deployment strategy to optimize solution performance based on the unique characteristics of the users and data and the infrastructure.
* Peer-to-peer and loosely coupled
With SOA, there is no master service, so every service is capable of complete fault tolerance. Any request can be handled by any service of the same type on any server in the system.
With no single point of failure, IT can deploy BI with confidence. This type of reliability builds the trust the user community needs for broad adoption.
These peer-to-peer services are loosely coupled, so no service needs to know how the others do their jobs, what they are doing at any time, or even where they reside. Business logic is kept separate from underlying infrastructure, so many data sources can easily be integrated.
For example, the BI tools of organizations that have identified SOA as a key requirement operate in a peer-to-peer mode, so their IT function has the flexibility it needs in deciding how services are distributed. Organizations that wants to manage reliability during peak periods of usage can leverage their development and test environments for those few days a month when they need to accommodate the increased demand for processing power. As a result, they can meet their service-level commitments without having to invest in additional hardware.
* Interface-based and coarsely grained
SOA is interface-based, so common messaging enables communications between services. Services interact using SOAP and XML to exchange data, parameters and processed results. For BI, the availability of the right communication protocols and open APIs is increasingly important because of the need to deliver BI to everything from mobile devices to enterprise search tools and other applications.
Finally, SOA is a good fit for BI because the services are coarsely grained, that is, defined at the business level rather than the application level. This reduces network traffic and simplifies integration. Coarsely grained services enable the reuse of these services across processes and applications, which is critical to efficient deployment of solutions.
Coarsely grained services are especially important for BI solutions considering the range of methods used to access information. A user needs to be able to quickly derive the same answer to a key business question whether that information is coming from a standard report or an ad hoc query, and regardless of whether the information was accessed through the BI application, from a mobile device or from within another application.
Key to achieving this in an efficient manner is using a common set of services built at the right level so there isn’t too much chatter between services to increase lag time.
Coarsely grained services are also critical for enterprise BI to ensure solution agility. To deliver the wide range of BI capabilities across a variety of environments and reach all users requires an architecture that quickly adapts to change.
With a common set of coarsely grained services for all capabilities that have been purpose-built for a BI function, a solution can evolve quickly to new requirements. Once again, this is especially important when looking at a tool that needs to reach users who work in environments outside the traditional BI realm. The right SOA foundation ensures new technologies can quickly be made available to all capabilities, instead of introduced as capability-specific solutions that need to be built and maintained independently.
Organizations should insist on enterprise BI solutions based on SOA as a prerequisite for success. Using SOA’s single set of standards-based, peer-to peer, purpose-built services and single, open API, lets IT take advantage of existing platforms, operating systems, hardware and security controls. This openness also means fewer moving parts to deploy and maintain, making for an efficient, reliable and agile performance solution across the enterprise.
For IT, this efficiency translates into ease of deployment, maintenance and change. For businesses, users can access data sources on a variety of platforms without inevitable infrastructure changes impacting their ability to obtain the information needed. The solution is up and running faster, business users have access to more information and they are not inconvenienced by back-end changes.
Hanniman is a senior product marketing manager at Cognos, an IBM company. She can be reached at jennifer.hanniman@ca.ibm.com.
