SQL Server Not Best Ingredient In IBM i BI Pie

February 18, 2013 Dan Burger

What does business intelligence and analytics mean to you? There are many answers to that single question, but information delivery is often the common denominator. Most people who think of BI think of dashboards. The capability to produce Web-based reports is likely to be the guiding light for most BI projects in the majority of IBM midrange shops where the dissemination of useful data, primarily residing in the DB2 for i database, is both inefficient and inconvenient. How do you fix that?

It is not with a green-screen query tool designed for last century’s business objectives, that is for sure. But it is because of that tool–the much-loved, much-maligned Query/400–that shops are moving in the direction of Microsoft‘s SQL Server database and the ubiquitous Excel spreadsheet. The familiar format of Excel is a lure, but the “we know how to do it” attitude of the SQL Server staff is what really seals the deal. Problem solved or problem just kicked down the road a bit farther?

Let’s take into account more than just the enthusiasm of the SQL Server advocates. Enthusiasm is an admirable quality, but if it involves charging down the wrong road, or blocking a well-thought out plan, it can do more harm than good. It’s worth the time to study the options. And that begins with realizing there are options.

Much of the published discussion about business intelligence and analytics places the focus on multiple databases and environments where the integration, security, and management puzzle becomes increasingly difficult. That may be where you need to go, but the majority of IBM i shops are not looking for that level of complexity, cost, and deployment time. (Beware of marketing messages that promise “deployment in less time than it takes to hard boil an egg.” They may deliver a simple solution, and that’s fine if that’s what you want, but realize BI sophistication is not cheap or easy.)

IBM has a strong i-centric customer base in vertical industries such as financial services, hospitals and pharmaceuticals, Web commerce, and tier one retailing where BI and analytics are taken to the highest levels, but the majority of its customer base carves out business in the SMB and in B2B industrial, distribution, and manufacturing where the most likely BI projects are not nearly as tough a nut to crack.

For the most part, IBM concentrates its efforts on the enterprise where technology acquisitions such as Cognos in 2007, the OEM deal with Information Builders that same year (which resulted in Web Query), and the purchase of SPSS in 2009. These moves were either aimed at introducing the high end of the IBM i base (as well as mainframe and AIX customers) to IBM branded Web-based business intelligence or you could say the adoption rate just happened to run in that direction. (Not likely.) Regardless, the majority of IBM i users have not made use of either Cognos, SPSS, or Web Query.

Just a few weeks ago, I talked with Doug Mack, a DB2 for i business intelligence consultant within the DB2 Center for Excellence team, about Web Query. Much of the article that came from that conversation focused on the complexity issue as it relates to the SMB customers and their less than warm reception of the product, which is delivered in a starter version (minimal user licensing) as part of the IBM i operating system.

Two companies that also have a stake in the IBM i business intelligence software business are New Generation Software and SEQUEL Software. Both companies are making their way in the IBM i small and midsize customer base by offering BI and analytics software that is less costly, less complex, and with a shorter deployment time than the products available from IBM.

Bill Langston, marketing director at New Generation, and Mike Stegman, senior data access consultant at SEQUEL, provided views of the BI on IBM i playing field during a telephone interview with IT Jungle last week. Both expressed the opinion that too many business intelligence projects in IBM i shops are being driven from the Microsoft side.

Part of the reason is that so many of the small and midsize IBM shops are still reliant on Query/400, a tool that is limited in its functionality and tied to a green-screen presentation. In BI planning discussions (if there are even discussions that involve the IBM i), the green-screen is a curse.

“[Decision makers] equate the limitations of the IBM i with the green screens,” Stegman says. “It’s often the end of story until they are shown otherwise.” If no one explains that the IBM i has non-green-screen options, the SQL Server backers win with little more discussion about BI topics. However, it can also be pointed out that Query/400 requires IT technical skills and involvement, while modern Web-based BI is set up so the end user operates without technical skills, which allows IT to step away after the set-up, deployment, and initial hand-holding is completed.

When the only Web-based business intelligence software option being presented is the SQL Server and Excel combo, it is going to win those battles, just like most comparisons that pit green screens against GUIs. Inconspicuous in the BI discussions are the GUI and browser-based tools designed for the IBM i. Why are they unnoticed? Lack of modern IBM i skills on the small IT staffs and a minimal understanding of the platform dooms many BI on IBM i projects.

The people with SQL Server skills will obviously say BI projects work best when the data is taken off the AS/400, as they call it. (They prefer calling it the AS/400 because that name perpetuates the idea of old technology among those who know little about the platform.) However, there are solid reasons why sucking the data from the DB2 database is a bad idea. One way to sabotage the success of a BI project is to inject IT personnel who are unfamiliar with the DB2 for i database.

“Business intelligence begins with data and the people who understand the data are the best ones to determine how to get at the data and organize it,” Langston says. “They know the subtleties of how the data was organized. The farther away you get from the origin of the data, the more difficult the BI project becomes. There is still a lot of value in the people with skills who understand the business and the data.”

Articles that emphasize the importance of database skills have been published in this newsletter in the past. Get Database Skills for Career ROI puts forth the idea that these skills will become increasing sought as more the amount of data increases and the data mining intensifies. Those who have the enterprise software background, knowledge of the IBM i and its database bring a lot to the table.

Minimal staffing levels are a big issue in the SMB. They are established to maintain the status quo rather than to lead new initiatives like BI and analytics. You have to like eating the same thing for lunch year after year to make a status quo plan. And you have to realize a much broader skill set is required to move forward compared to when a simple 5250 interface was perfect. Consider, for example, how many layers of technology there is to deal with that involve smartphones, tablets, PCs, and different releases of operating systems for the various pieces.

IBM is only recently getting the message to customers that they need to improve knowledge of DB2 and its modern capabilities. Companies aren’t staffed for it on the IBM i side, but both Langston and Stegman noted there are people on the Windows side ready to take on these responsibilities in many cases.

Stegman points out that when the main application data is handled by the IBM i, and there’s a need to integrate that data with data from other systems, a problem can arise with multiple systems handling the same information and getting different results (different versions of the truth, as it is commonly called). It is more likely to happen when the users of various smaller databases (customer resource management and employee time management are two examples) are working independently and going to the larger system (the IBM i) and grabbing info. His recommendation is to consolidate reports on the main IBM i system (scalability, reliability, and security are its strengths), which is grabbing smaller bits of data from the various databases. SEQUEL Software, as Stegman well knows, provides the graphical (not green screen) tools to do this with drill-down capabilities and a Web interface delivers browser access that users are often led to believe only exists on the Windows side.

In some cases, the awareness includes options like NGS-IQ and IBM Web Query. In those cases there is much to compare with what the SQL Server proponents are backing.

Integration, as you might expect, does add some complexity. Eventually companies will fight that battle, but for many of the IBM i SMB shops modern browser-based BI is first focused on getting data from DB2 to the people who can make use of it. They may also find that some work on the DB2 for i database is required to move forward with BI planning and deployment. Another warning for BI projects that are driven by teams without IBM i knowledge or business application knowledge is the assumption is that all these things will be well understood. Portions of a BI project can be simple, but the more that’s being demanded the less likely it can be done overnight. A messy database, which often is not noticed until a project like this comes along, takes more work to make the data useful.

For many of the reasons touched on in this article, Langston sees the cloud playing a larger role in the future of BI in small to midsize shops. And New Generation has put that idea into the marketplace using a private cloud in combo with a managed business intelligence service.

“We’re just taking baby steps right now,” Langston says. “We understand that access to data for mobile and Web users isn’t something that companies want on their production machine–they want it segregated–so we are working to help with that.” He expects most customers will have on-premise BI, but will complement that with some cloud-based BI done off-site that would support Web and mobile users.

IBM i shops have investments to consider and in terms of business intelligence and analytics, there are options that require more attention than the first guy who stands up in a meeting and says, “I can do that.”