Education the First Step in TEMBO’s Quest for Database Modernization
April 9, 2013 Alex Woodie
It’s been nearly a year since TEMBO Technology Lab launched its flagship database modernization tool, called Adsero Optima Foundation, in the US market. Since then, the product has garnered interest from dozens of American IBM i shops, who are intrigued with the prospect of migrating from old DDS to the SQL engine. While the folks at TEMBO would have preferred a warmer reception–and a quicker sales ramp-up–they understand that there’s a need for greater education on the topic of database modernization.
Make no mistake about it: TEMBO is in the database modernization business to make money. It is true that the South African software company and its master distributor for the Americas, Southern California’s inFORM Decisions, have noble aims. They are devout adherents to the idea that the IBM i server is the best business platform in existence, and that its users will be best served by investing in it, not relegating it to the scrapheap. So if one believes that the IBM i platform is at risk–which it most assuredly is–then why not make it one’s business to help it?
While it’s not hard to get a consensus that the IBM i server is endangered, it is difficult to get a consensus on the cause of that threat. To be sure, the 5250 green-screen interface doesn’t help. The higher upfront price of most IBM i servers compared to Wintel servers undoubtedly discourages further tire kicking by potential customers. Legitimate innovation and progress is being made in the ERP stacks of Microsoft, Oracle, and SAP in the areas of cloud, big data, and mobile. Then there’s the whole “legacy” thing. “The AS/400–Isn’t that what grandpa used to work on?”
Grandpa’s Coding Practices
Certainly these are all challenges. But TEMBO CEO Marinus van Sandwyk has a different view on the biggest threat to the IBM i platform, one born from conversations with actual programmers and CIOs, not technology fashionistas and accountants.
“The single biggest reason people move away from the IBM i platform is not due to functionality. It’s due to the fact that they can’t respond fast enough to changes in the business environment,” van Sandwyk tells IT Jungle. “The functionality that it provides them is dramatic. It provides them with a competitive advantage. But the reason why the IBM i server is perceived as old is because customers can’t respond fast to changes in the business environment due to the way they used to code those systems.”
The answer, of course, is to change how IBM i shops code their systems, which, of course, is easier said that done. Instead of writing database validations directly into the RPG or COBOL program, these validations rules should be pulled out and housed separately. Instead of using the DDS approach to deliver “native” I/O, IBM i shops should house the validation rules using SQL.
Aside from the fact that the rest of the modern computing world uses SQL as the interface between their applications and databases, there are other big benefits to this approach.
The biggest benefit that IBM i shops will get from centralizing their database validations is that maintenance activities are dramatically reduced. Currently, whenever a major change is required, IBM i shops must comb through all their RPG or COBOL programs to make sure that all the locations that contain validation rules are updated appropriately. This approach is time consuming and a major impendence to productivity.
None of this is news, of course. For years, IBM has been telling IBM i shops that they are best served by adopting the SQL engine for some types of processing, especially sequential processing and batch runs, which normally see big improvements by moving to the SQL Query Engine (SQE). (The old DDS approach often yields better performance for random access workloads.) Many new programs written today do use SQL. The reason older applications haven’t been retrofitted to use SQE is that it would be tremendously invasive.
What is news is that TEMBO claims it has achieved a breakthrough that reduces this risk and invasiveness by automating most of the migration process. AO Foundation works by extracting metadata from the existing DDS-driven database and applying that data in the new SQE-driven database; no recompiles are necessary. The two databases are kept in sync until tests are complete, and then production is switched over to the SQE database.
Strength or Weakness?
The local South African market was the proving grounds for AO Foundation, and when van Sandwyk was satisfied that the product worked as advertised, he made his way to the US. TEMBO made a big splash at last year’s COMMON conference in Disneyland, and announced its partnership with inFORM Decisions in the fall.
With a solid year of experience in the American market under his belt, van Sandwyk has learned a few things. For starters, the learning process could use some attention.
“When I got to the US, and found that we actually had to explain to the bulk of the people why they had to move to SQL engine, I was totally blown away,” he says. “We then realized we had an educational process on our hands, which means it’s going to take longer.”
It turns out that many American IBM i shops believed the native I/O enabled by DDS was a strength. After all, it meant that you didn’t need a dedicated database administrator (DBA), like SQL Server and Oracle shops did. But instead of enabling shops to be lean and mean, it turns out they’re just loading their programmers with the sort of tedious maintenance tasks that are unheard of in SQL Server and Oracle shops. A perceived strength turns out to be a weakness, which is never fun to hear.
“You will see that Rochester has actually changed their position on that,” van Sandwyk says about the need for DBAs. “The reason for that is simply that 80 percent of the functions–the database relationships and the database validations–should reside with the database engine, and the only way that you can truly manage that sensibly is to have a solid database engineer on your staff.”
While the ramp-up in the US is slower than expected, it’s making progress. TEMBO doesn’t have that critical first reference customer in the US yet, but has a handful that are close to making commitments to use TEMBO technology for their database modernization projects, and another 50 or so in the pipeline, van Sandwyk says.
TEMBO plans to spread the message of database modernization through several webinars, user conference appearances, and magazine articles. Van Sandwyk, who recently traveled to IBM’s Dubai facility to discuss database modernization and IBM i on PureSystems, plans to travel to the US next month to meet with prospects.
Interestingly, PureSystems are key to TEMBO’s message, because they can deliver operational savings to multi-platform IT staffs, and provide an incentive to do the work to modernize their IBM i apps to keep them viable for another 10 years or more. “We’re doing a lot of work with the Pure Systems team at the moment,” van Sandwyk says. “The PureSystems people understand what we’re doing, and are very excited.”
The diversion into education was not expected, but TEMBO realizes it’s at the beginning of marathon, not a sprint, and a good foundation is critical to future success. That doesn’t mean there isn’t a sense of urgency. “We must help with the education process and convince the people that it’s not a ‘nice to have’ to move to the SQL engine,” van Sandwyk says.
TEMBO has other products in its long-term roadmap–think grid computing on self-managing nodes, which goes along with van Sandwyk’s expertise in OS/400 clustering and interest in PureSystems. But these future products, and future TEMBO profits, are all dependent on getting IBM i shops to take that first big step onto the SQL engine.
“If people don’t commit to move to the SQL engine, in our opinion, they’re wasting their time, because any move they do from a modernization perspective is tactical at best,” he says. “It may buy them a year or two years. But fundamentally they have not actually addressed the issues that are creating the perception that the applications are old. It’s an absolute imperative. If they don’t move, they will lose the platform.”