ARCAD Software Takes On Database Modernization
November 4, 2014 Dan Burger
If you’ve seen IBM‘s Redbook on modernization for i-centric customers, it’s clear the direction Big Blue recommends includes the DB2 for i database as an integral part of the modernization process. In application-first environments, which dominate in IBM i shops, database modernization is, at most, an afterthought. That’s changing, and ARCAD Software, known for its application change management software, is changing, too. Its latest release covers DDS to DDL conversion, data integrity checking, and field renaming.
Philippe Magne, CEO and chairman of ARCAD, explains the rationale behind this extension of enterprise modernization capabilities. “It is the DB2 database that has given IBM i its unrivalled reputation for reliability, security, and scalability, and it will continue to be a major development advantage of the platform–on condition of having a fully relational SQL database. SQL is not only the de facto standard, but it also scales well to large quantities of data and offers industry-beating performance on IBM i,” he states. “With SQL, you render data architecture readable, including migration toward explicit field names and links between tables. You also open up the database for update to all external tools and applications, with no risk to its integrity.”
IBM’s investment in the IBM i operating system releases and Technology Refreshes has primarily focused on the database, and enhancements there have been DDL (SQL)-based since IBM i 6.1. (RPG Open Access and free-form RPG are the notable exceptions.) As everyone should know by now, IBM has steered DB2 for i in the direction of SQL for the same reasons it has modernized RPG–to make it more mainstream with the modern multi-platform, multi-device IT world.
There are signs based of increased interest in sessions at the technical conferences that the use of SQL is rising as the IT leaders are recognizing the power it brings to Web, mobile, and social projects. In many of those circumstances–particularly when flat files or hierarchical designs are still in use–the current DDS-based databases will not support the application modernization objectives. The blame for that can be pinned on a lack of interoperability with the most popular Web and mobile technologies.
“To fully exploit the underlying IBM i platform, you need a three-tier application with a genuine relational database, stored procedures, and a graphical interface–and optionally, Web services. This requires a staged modernization of both the application and database layer,” Magne says.
ARCAD’s Transformer DB suite fits into the ARCAD Transformer product line, which in its latest update is referred to as V10.04. Other standalone products in that product line include ARCAD Transformer RPG, for conversion from RPGLE to free format RPG; ARCAD Transformer Field, for field size expansion; and ARCAD Transformer Unicode, used to “internationalize” IBM i applications (Asia, in particular).
Transformer DB is designed to aid database migrations by providing database integrity checking, conversion of DDS to DDL (SQL), and field renaming. Data integrity is important when transitioning from the DDS model to DDL, because the data management technique is different in the two models. To anticipate errors, Transformer DB searches for and analyses incorrect data and also identifies its cause.
The limited length of field names in DDS (and in RPG) is often unworkable in a modern application. The Transformer DB field renaming feature enables developers to “refactor” source code and file definitions and automate the mass changes involved.
Also in the ARCAD software arsenal is a separate tool called Observer, which can also assist in understanding the existing database relationship model, even when flat files are part of the data model. It does this using data analysis engines that drill down into program code to “understand” the links between fields and propose a relevant structure.
Magne believes managing change and mitigating risk are important in the modernization process.
“Even with the level of automation we provide, modernization of the application and database involves large-scale changes in source code and data,” he says. “And a high level of change means a high risk of regression. Our customers usually need to keep their businesses running on one application version while modernizing in parallel on a separate branch. This is why we stress the importance of using lifecycle management tools, and especially test automation, in a modernization project. It is the safest way to ensure the reliability of your applications and minimize the impact on your end-users.”
Many companies still unfamiliar with the database modernization wonder about the return on investment. Several factors should be taken into consideration when considering ROI. Data volume, for most organizations, is growing. SQL scales much better and, depending on how I/O bound your program is, converting DDS to DDL can give some programs an immediate performance boost. SQL isolates programs from database structure changes far better, which eliminates mass compilations. Also, when database skills are taken into account, SQL skills are much easier to locate and add to your staff.
You may have noticed database modernization has yet to be a widely accepted practice in the IBM i community. In part, that can be attributed to thinking that it’s a proven entity, thoughts about integrating with other platforms have yet to be formulated, and modernization projects have been framed from an application development perspective.
“The increased initiatives for Web- and mobile-centric applications continue to drive the understanding that enterprise modernization is much more, however, than just the screens and devices,” notes Floyd DelMuro, ARCAD’s vice president of sales in the Americas. DelMuro believes the real measure of success for modernized applications is determined by re-engineering the data layer.
“There is a need for increased database performance and versatility, advanced query and business analytics that are challenging to the databases of old,” DelMuro says. “I am seeing the elimination of batch-processed applications, or a requirement for improving batch-processed applications, to have applications that are more flexible and responsive. I am also seeing companies expanding into new markets where language requirements are forcing changes in data capture and requirements.
“This need to re-engineer the data layer has increased the need to manage change and mitigate the risk in the modernization process,” DelMuro continues. “In today’s highly competitive economic environment, the company that can automate these large-scale changes in source code and data will have the advantage.”
The IBM Redbook “Modernize IBM i Applications from the Database up to the User Interface and Everything in Between” is available as a download here.