Thrash and Trash: Moving Away from Dependence on Legacy Systems with Object-Oriented Interfaces
Articles and Tips: tip
Novell Systems Research
01 Feb 1996
In the Novell Systems Research Department, we often find ourselves using a legacy system in our lab experiments. This may seem strange, since we work with some of the newest equipment in the industry--leading-edge equipment that makes last year's or even last month's hardware seem antiquated. This constant turnover of test equipment creates an archiving issue for hardware, software, and applications.
The systems we are using today are not the same systems we will be using in two years. For instance, in the past two years we have moved our imaging performance testing from Magneto-Optical (MO) media to hard disk. The reasons are more for performance than for economy. With 1992-93 implementations of image stores and the availability of an HP optical jukebox as part of Novell's MSS, we found ourselves using MO media. In fact, we are still using a Panasonic MO jukebox, along with an imaging applications from PaperWise and File ABC. These have served us well through a myriad of grueling tests.
The jukebox and these applications will continue to be used in several long-term studies. However, the jukebox's basic function has changed. It is no longer an image store based on optical media. It is now being targeted for testing as an image and document archival device, where we store 3- to 5-year-old images that are in low-usage retention--not moderate or high usage.
The migration of our treasured jukebox farther and farther from the server is not a technological marvel, but rather an economic marvel. Two years ago at COMDEX one of the vendors, Micropolis, showed us conclusive proof that magnetic media would be less expensive and have higher capacity than MO media could provide. Considering the price of jukeboxes today, it seems this prediction has come true.
That information, combined with the performance and retrieval needs of most imaging systems, convinced us that the days of MO for image retrieval were numbered. Only large companies performing multiple, nonsequential image retrievals would continue to use the jukebox as part of the online direct-access services. CD-ROM technology like our Pioneer changers is better suited for moderate sequential use by a limited number of image retrievers.
Today optical media and jukeboxes are moving toward the back-end of imaging and document management systems. In fact, we see fewer MO autochangers for imaging in the future, and more hard disk and CD-ROM technology. Whereas in the past we used the jukebox as the place to store image for immediate retrieval, today we are looking at the jukebox as an archival and storage retention device.
Bridging the Gap
What does all this hardware talk have to do with object-oriented interfaces? In Systems Research, we have been using AppWare as an object integration tool for our image testing, and we're currently looking at Borland's Delphi for comparison. There are numerous papers in the industry on Obect Orientation (OO) that give differing perspectives on where the industry is headed. Because of our success in objectizing components in our imaging test bench, we offer the following information about OO and legacy system integration.
As mentioned above, we are using AppWare and independently-developed objects to integrate various components (front-end and back-end) such as our Pioneer CD-ROM autochanger for CD integration testing, a TMS image viewer, and several database engines including Oracle and Sybase. While the back-end technology is changing and the implementation methods vary, we are finding great comfort in being able to manipulate objects on our test bench, recompile (in minutes), and then retest with a different image store, viewer, or database. And it's easier than we ever dreamed possible!
These benefits alone have prompted us to consider OO front-ends and legacy systems as a natural and normal evolution of our testing systems. You might find these methods equally applicable as your production systems evolve.
* Originally published in Novell AppNotes
Disclaimer
The origin of this information may be internal or external to Novell. While Novell makes all reasonable efforts to verify this information, Novell does not make explicit or implied claims to its validity.