Analyst Study Says Plato Has a Lot to Learn

When the philosopher Plato said, "Knowledge is true opinion," he had no idea that a company calling itself Microsoft Corp. would one day code-name a piece of its SQL Server 7.0 database after him. Or that the product would be designed to contain and manage knowledge -- albeit in the form of data rather than opinion.

Microsoft is trying to stake a claim in the enterprise-class database world with SQL Server 7.0, and the OLAP Services feature the company referred to as Plato during the software’s development.

With the release of SQL Server 7.0, Microsoft is pushing to challenge the classic databases from companies such as Oracle Corp., IBM Corp., and Informix Corp. (

A report titled "Is Microsoft Plato the Best Philosophy for BI?" by market research firm The Robert Frances Group Inc. (RFG,, states that SQL Server 7.0 regrades the playing field for data warehousing into a more competitive one.

The report, however, demonstrates that SQL Server 7.0 is not on the same level as its competitors yet, particularly when it comes to data warehousing. A major factor is the OLAP Services.

Cal Braunstein, director of research and CEO at RFG, says SQL Server 7.0 is not as complete a product as its competitors.

"The high-end functionality that other tools have is not available in Microsoft’s OLAP services, so users have to buy third-party tools to get it," he says.

Microsoft also needs to overcome scalability issues facing SQL Server 7.0. First and foremost, other databases are closer to achieving 24x7 availability. That is becoming a must for e-commerce sites.

Second, SQL Server needs to be able to handle more concurrent users than it now supports. According to RFG, 50 users is about the maximum number that can simultaneously access the database without negatively impacting performance. "I wouldn’t call that enterprise-enabled," Braunstein says.

Adam Braunstein, a research analyst at RFG, points out that companies can achieve similar enterprise-class performance by stringing together many SQL Server boxes and implementing fail-over clustering to ensure that even if one box goes down, the system still runs.

This is the approach that major Web sites that run NT use. They string a bunch of boxes together, and only use each box at about 30 percent of the CPU’s capacity, which also helps to ensure that the machines won’t crash.

Such a solution, however, can easily require enough servers to drive the cost up into the range of buying a bigger box with Oracle or Hyperion OLAP software installed.

An important factor for improving scalability on both of the aforementioned fronts is improving the reliability of the operating system that SQL Server runs on.

Windows 2000 will improve scalability, as will the faster processors coming out of Intel Corp. But RFG says these alone are not enough to bring SQL Server up to par with its competitors.

Not surprisingly, RFG’s report answers its initial question: Is Microsoft’s Plato the Best Philosophy for BI? The answer is a bit of an oxymoron: No and Yes.

For companies needing a data warehouse with considerably more than a terabyte of storage, thousands of concurrent users, 24x7 uptime, and OLAP services that do not rely on third-party tools, RFG does not feel SQL Server 7.0 is up to the task.

Companies that have SQL Server in place and are using Windows NT and Office as major parts of their business infrastructure, though, are better off staying with SQL Server.

Bearing in mind that SQL Server and its OLAP Services will improve over time, companies need to be willing to grow with the product.

"If you’ve got SQL Server and it’s working without costing too much, there’s probably no really good advantage to switching -- unless you have an imperative business reason," Cal Braunstein says.