The Data-Centric Depiction Of IBM i
January 18, 2016 Dan Burger
You have data. You’ve been collecting it in pretty much the same fashion for years. The only thing that has changed is you have more of it. But are you getting more out of it? I’m not talking about data analytics. The world doesn’t need another person talking about data analytics. This is about improving business information integration and the role IBM‘s DB2 for i has played and will play in the future.
Data-centric development is not a new concept. It hovers in the background of the IBM i roadmap, but is maybe the most critical element to the future of the platform. When people ask whether IBM is investing in IBM i, the database is at the top of the list of answers. This is why it is not the AS/400 anymore. This is one of the big reasons why it is a modern, more efficient system.
What is new, according to Mike Cain, a senior technical staff member at the DB2 for i Center of Excellence, is an increased emphasis on communicating the data-centric message. How to begin data-centric development and how to understand the process from start to finish needs some work. Cain’s already taken this new approach to the worldwide IBM i community with a series of DB2 technical forums that have taken place in Europe, Asia, Central and South America. He’s preparing to do the same with the database track at the RPG & DB2 Summit that’s making its next appearance in March in Dallas, Texas.
The simple and straight-forward description for data-centric development is that it allows the database to do more work keeping track of and securing the data. The database is the key and it is capable of providing many more advantages and efficiencies than it has been tasked to do in the past.
There’s not a lot of data-centric development going on, much to the dismay of Cain and the DB2 for i database team. On the plus side, SQL use is widespread. Database enhancements are the focus of IBM’s Technology Refresh program. And awareness about database modernization has increased. So it seems the message has been delivered and read. There’s little evidence, however, that data-centric development is building momentum.
One reason for that is that knowledge of the database is a new frontier. Many IBM i shops operate without database administrators or database engineers. Database modeling is an important factor in building successful data-centric applications, but database modeling skills are a rare commodity. In fact, many database skills–such as data manipulation, set-at-a-time programming, and data definition languageâ€”are hard to find. Tools like DB2 Data Migrator and Data Studio that play important roles in data-centric programming are largely unknown.
“Most people accept the message that the database is important and [business success] is all about the data. The use of SQL is accepted. People are working on the proper use of SQL, on doing more with their data, and on honing their craft,” Cain says. “Some people accept the premise, but do not have a full grasp of how a concept like data-centric development works in the correct context. We continue to bang the drum about proper data-centric development. We explain the technology, features, functions, benefits, but we also demonstrate how it gets applied and how it makes sense by using realistic examples that goes from design to build to secure to tune. What to expect at the start and at the finish and what will be seen along the way.”
A better way to learn the data-centric development process, Cain says, is to have a real world model to not only talk about, but to also demonstrate how queries are run and results are achieved. He’s added that to database training sessions he and his team present. He also gets the point across by introducing the role of a database engineer, the importance of organizational support, and the development of skills and designation of responsibilities.
“We are trying to explain it better, to make it seem easier to get started with less risk,” Cain says.
“In the old days, people did data modeling by the seat of their pants, on the back of a napkin, or ad hoc as they wrote their DDS,” Cain notes. “It was all hand work.”
Using Data Studio allows data modeling to be done graphically instead of using only words and phrases. Certain programming steps are repeatable, and Data Studio streamlines this. It can also provide a warning when table-to-table or object-to-object mistakes are made. It’s much easier to manipulate the data model, Cain notes.
Most people Cain talks with don’t know Data Studio exists. They are still doing DDS with 80-column, 5250 green-screen tools. They’re surprised to learn about Data Studio, that it interacts with IBM i, and that it’s free.
An intro to Data Studio is one of the 14 database sessions Cain and his colleagues Tom McKinley and Doug Mack have on the Summit agenda. The session list includes instruction on data definition, data manipulation, database design and modeling, SQL procedural language, DB2 SQL performance tools, among others. The entire list, with session abstracts, can be found here.
Cain and the other members of the DB2 for i Lab Services team that are called on to assist companies with their efforts to establish data-centric development all use Data Studio. The tool is free and easily downloaded.
Because the tool is free, it removes the “approval obstacle,” Cain says. It allows users to be productive, but it’s not a “starter kit.’ A lot of clients only use this for data centric development. Data Studio only does physical modeling, not conceptual models and logical models that are used in large enterprises. Physical modeling is all most organizations need, Cain says.
Before downloading Data Studio, there are a few words of advice.
“You need a foundation based on data modeling knowledge and skill. Data Studio is tooling for data modeling and implementing the data model. As part of the foundation, you need to know about primary keys, parent-child relationships, or how to describe the attribute of a column,” Cain says. Without the foundation knowledge, Data Studio won’t make sense. After attending data modeling and Data Studio sessions, data modeling is easy to do with Data Studio. People have been doing rudimentary data modeling for years, but those who do it using best practices and with an understanding of relational data modeling are rare.”
Cain has a long relationship with the twice-annual RPG & DB2 Summit conference that is run by System i Developer, an educational endeavor headed by Susan Gantner, Jon Paris, and Paul Tuohy.
“Like me, I suspect that many RPGers would love to embrace SQL and set-at-a-time processing,” Gantner says. “What slows us down is that most existing database designs are not optimal for that, and our one-record-at-a-time habit has been with us a long time! Many try to keep up with new DB2 for i capabilities and best practices in database design and usage over time. But I often feel that I’ve lost the thread that ties all of it together.
“That’s what I liked so much about the idea of a soup-to-nuts database track. It starts with data modeling and takes us through the whole process, including discussions on role-based security and why we need to become database “engineers.” Mike promises to help us find that thread to pull it all together and complete the picture of how to design it, build it, secure it and tune it.”
The RPG & DB2 Summit is scheduled for March 22-24 in Dallas, Texas. Details and registration information can be found at this link.