Deployment options increase for meeting BI performance challenges

Learn how different business intelligence deployment options can affect BI performance at small and medium-sized businesses (SMBs) involved in SQL Server BI projects.

As an increasing number of companies look to implement business intelligence (BI) systems, new ways of meeting their BI performance and scalability needs are emerging, in the process giving potential technology buyers additional deployment options to evaluate and consider.

At a basic level, most organizations -- especially those less-than-gigantic companies often classified as small and medium-sized businesses (SMBs) -- have three very simple goals for their BI systems:

  • They have to provide reliable answers to business questions, meeting the overall promise of BI itself.
  • They have to perform queries quickly while also enabling users to interactively drill through data sets as much as possible.
  • They have to be easy to deploy, without years-long planning, design and implementation cycles.

More specific capabilities differ from company to company. Everyone wants dashboards and reports, but what people want to see in them varies. Those kinds of features are the major points on which BI system vendors compete. But how can you ensure that the underlying performance and deployment issues are addressed as part of a BI project?

One way is to focus on the combination of BI systems and database engines. Microsoft SQL Server, for example, includes an Analysis Services component that is specifically designed to handle BI workloads, and it’s entirely possible to build both custom and packaged BI systems on top of that platform.

Better system design begets better BI performance 

There are some advantages to such approaches. The platform vendor (Microsoft, in this example) often builds up significant experience in BI system design over time through all of its dealings with customers and BI vendors. That shared experience can lead to better system designs, and better-designed BI systems are better-performing BI systems.

Microsoft, to continue using it as an example, has a wealth of publicly accessible best-practices information and system design patterns, along with in-the-trenches studies on how to wring the best performance from BI systems built on SQL Server Analysis Services.

Another potential option is a Software as a Service (SaaS) approach to BI. Just as companies are moving email, customer relationship management (CRM) and other IT functionality into the cloud, you can purchase cloud-based BI technology from a growing list of vendors.

SaaS BI systems offer the promise of rapid deployments along with access to a back-end architecture that most SMBs couldn’t afford to build out on their own. That creates the potential for Fortune 50-class performance -- but the cost of the SaaS BI architecture is divided across numerous customers, making the capital and operational investment more affordable for individual organizations.

Internal hurdles for SMBs on BI performance, functionality 

As mentioned above, system design is the prime driver of business intelligence performance. However, few SMBs can afford to employ a full-time BI design expert. That means homegrown BI systems are often built for them by consultants, who may or may not fully understand what a company needs from its BI system. They also may or may not be around later to tweak the design to accommodate ongoing changes in the business environment or correct initial design errors or misconceptions.

Packaged BI systems seek to avoid that problem by creating preconfigured product offerings that are designed around common business concepts and practices, instead of requiring a customized system for each customer. At a certain level, this out-of-the-box approach can work well. After all, most companies use packaged software for email, accounting, human resources, enterprise resource planning, CRM and other functions. Because BI systems draw much of their data from those kinds of operational systems, it stands to reason that -- for certain companies -- packaged BI tools will meet business needs.

Such systems may be designed and tweaked by BI vendors or systems integrators for optimal performance on SQL Server or another chosen database platform. Although the black-box nature of preconfigured systems makes it hard for users to modify them for increased BI performance, they often are easier and less expensive to implement than built-from-scratch systems are, helping to address the deployment concerns that many SMBs have about BI technology.

The good news is that with the rapid growth of BI as a product category, businesses now have more choices than ever for obtaining or building a high-performance BI system. Solutions to the BI performance challenge are out there -- you just need to find the right one for your organization.


Don Jones is a senior partner and principal technologist for strategic IT consulting firm Concentrated Technology LLC.

Dig Deeper on SQL Server Business Intelligence Strategies