DB2 & IMS

In the beginning, there was data, and data was good. But the data began to grow and become unwieldy. The data needed a form, and IMS was developed. The origins of IMS date back to 1966, when 12 members of an IBM team working with 10 members from North American Rockwell and three members from Caterpillar Tractor started the design and development of the Information Control System (ICS) and Data Language/I (DL/I). During the design and development process, the IBM team was moved to Los Angeles and increased to 21 members. This team completed and shipped the first release of ICS. In April 1968, ICS was installed. The first “IMS READY” message was displayed on an IBM 2740 terminal at the Rockwell Space Division in Downey, CA, on Aug. 14, 1968. ICS was renamed Information Management System/360 in 1969.

With the advent of IMS, data could now be stored in a logical, hierarchical structure. Applications could access the data in an orderly manner. For many years, IMS data was happy and content in full-function databases. As the volume of data grew, and as new applications required different types of access, other types of IMS databases, such as Fast Path and High Availability Large Database (HALDB), were developed.

More data meant more—and larger— databases. Databases could be up to 4GB with VSAM or 8GB with Overflow Sequential Access Method (OSAM), but hardly any were because storage devices were quite expensive. DBAs used compression routines to pack data as tightly as possible. They reorganized often to reclaim space. Databases were maintained during a large maintenance window. Basically, the entire weekend was free because no one did real business on the weekends. And the hours outside 8 a.m. to 5 p.m. local time also were available for processing.

In the beginning, IMS DBAs were plentiful. But when new technologies such as DB2 and distributed systems were developed, many IMS DBAs transferred to the newer technologies. IMS was considered the dinosaur of the data center—old technology that would soon be extinct. Even as IMS expertise dwindled, and nobody was being trained in IMS, companies realized the applications that used IMS were actually efficient and that it would make sense to keep those IMS databases around. In fact, the databases continued to grow.

Twenty years ago, time was plentiful, space was expensive, and it was easy to find IMS DBAs. The opposite is true now. What happened?

The Evolution

The ’90s changed data processing. Companies merged more than ever before, creating mega-companies with worldwide operations. The Internet gained popularity and showed businesses they could offer services in all geographies at all hours. The potential for generating more revenue seemed unlimited. Instead of selling insurance policies from 8 a.m. to 5 p.m. Monday through Friday, the insurance company could sell policies 24x7. The change was huge.

The upside was that companies could gain much more exposure and make more money. The downside was that business processes had to change— quickly. Let’s look at how an IT department had to evolve.

In 1972, for example, if you wanted to buy an insurance policy, you’d drive to an agent’s office during a weekday and fill out some paperwork. The agent would send that paperwork to the home office (allow three days for the mail to arrive). The data entry team would type your information into their mainframe (probably using punch cards) and a quote would be the output. Someone in the home office would print the quote and send it to your agent (allow three days for the mail). The agent would call you, and you’d drive back to the office and write a check for your insurance.

In 2008, when you want a new insurance policy, you log on to a Website, enter your information, and get a quote in a few minutes—even at 3 a.m. Simply answer a few questions, enter a credit card number, and you have an insurance policy.

5 Pages