Implementation of an Enterprise-Level Business System
Internet systems tend to have much more emphasis on the system servers, because the PC acts as a thin client and performs very little processing. Though, web-based applications tend to execute scripts on the client the majority of the application is executed on the host server.
Client server systems tend to be more complex because they typically require some client side software do be developed and installed. This can be tricky because the client software will have to run on unknown hardware and operating system. During the testing phase, the client software should be tested with as many possible scenarios as time will allow. The client software may need be compatible with Windows 2000 workstation on a Dell desktop and be compatible with Windows XP on an IBM Laptop.
Having a change control in place will help keep chaos from setting in. Change control will allow developers to see when a software bug was introduced into the system by tracking when changes were made to the code. Change control also allows the documentation group to track changes in their documents and rollback if needed. This process also involves assigning build number to software builds so they can also be tracked. This will assist in knowing what features are in what version of the software and keeping and archiving old documents, code and software builds.
Once a system is ready to put into place, the next job is to phase system in. The system change over can be classified into four different methods: direct cutover, parallel operation, pilot operation, and phased operation. Once the system is installed then any final customizations can be done before final acceptance testing.
What ever the implementation is, the important point is to have a well defined process. The