BI Driving Interest in Grid

Desire to minimize CPU dormancy indirectly benefits open source.

In thinking about strategic IT purchasing, high-level IT managers are considering what they'll invest in to support enterprise business strategy. Their decisions will likely revolve around the need for providing quick analysis of performance metrics, customer life-cycle management trends, and supply chain management data. The requirements for such business intelligence and service-oriented architectures will push data volumes to new highs. Consequently, those with greater than normal computing demands are exploring the potential performance/cost ratio of grid computing platforms.

Market researcher Evans Data Corp. recently released its semiannual study of development managers and application programmers in database environments. In the "Database Developer Survey" (available at, analyst Joe McKendrick covers a range of topics, including project planning and budgeting, BI, data warehousing, and operating systems.

"We found that grid is, in fact, growing quickly," says McKendrick, who notes 26 percent of those surveyed are considering implementations of grid, when only six months ago, it was 7 percent. In the June survey, Evans Data research found that 37 percent of 502 database developers are implementing or plan to implement a grid-computing architecture — an increase from 21 percent in a similar survey last December.

As IBM, Oracle, HP and Sun promulgate the economics of grid computing, they're attracting the attention of companies that want to leverage existing infrastructure rather than purchase new servers. By using otherwise dormant portions of infrastructure for spikes in computing needs, grid could be the answer to load balancing and failover requirements during peak computing periods.

A jump in interest around BI and analytics for business performance analysis was evident in the study. "Following the IT spending slowdown of the past few years, now 43 percent of companies are engaging in BI and analytic activities, as opposed to the 26 percent just a year ago," notes McKendrick, citing that as the trend in all industries. "They want snapshots across a range of areas to develop dashboards, through which executives see metrics on finance, sales, and supply chain issues." That's driving companies such as Business Objects and Cognos to promulgate dashboard concepts. "Vendors tended to be specialized with financial data, but they are expanding to address the breadth of organizations," he says, citing CRM as a prevalent area for BI.

That desire to analyze data with point-and-click functionality is driving interest in 53 percent of companies, according to the survey, which found about 44 percent of companies using BI for financial analysis and 35 percent using it for customer life-cycle management.

"That trend is important, as grid can be an enabler, as it makes the applications effective by handling growing volumes of data efficiently enough so that enterprise data warehouse environments are created more cost effectively," says McKendrick. He believes that companies with data warehouses scaling up to 20 terabytes and higher will see grid as a natural for mining huge data stores.

The question of the ultimate operating systems for grid computing also seems to be answered in the study. While Windows is strong because of its pervasiveness in the industry, Linux is becoming a "natural" for grid, according to McKendrick. Namely because IBM and Oracle are forging ahead of the pack with grid, Linux is well positioned to be the platform of choice. "For enterprises, support is key, so 'LAMPstack' will be the prevalent open source platform," he predicts, referring to the combination of Linux; Apache Web servers; mySQL; and PHP, Python, and PERL.

The building momentum for open source databases is evident in the fact that 64 percent of respondents in the study use them in some capacity. "That's up from just 7 percent three years ago," says McKendrick. The fact that 48 percent of that lot use MySQL demonstrates an affinity of MySQL for grid. While MySQL offers a cluster strategy with a highly available parallel server database, announced in April, MySQL is currently conservative in responding to hype around grid.

According to mySQL's Zack Urlocker, VP of marketing, "IBM and Oracle have great grid offerings, but they are not mainstream; they are specialized heavy-duty technology." Urlocker acknowledges that he does see an increase in interest around analytics from CIOs: "The need for more data means more databases, but handling that with limited budgets means limiting expansion license revenues." Right now, he points out, fully one-third of mainstream companies see no tangible reason to go to grid.

Susana Schwartz is a New York-based freelance writer specializing in emerging technologies and their impact on IT infrastructure.