UNIT 5: SYSTEM IMPLEMENTATION– WEEK 9
Implementation is the stage of the SDLC during which theory is turned into practice, that is, the system is eventually deployed, that is, put into use. The major activities involved are as follows:
- Documentation and final testing of all the programs required.
- Creation of all the master files required in the system.
- File Conversion
- Preparation of user and data processing department operating instructions.
- Education and training of all staff who will use the system.
- Change over strategies
- Commissioning of the new system.
When a new system is to be implemented, it is likely that the master files either do not exist, or, if they do, that they are not organised as required by the new system. Therefor before the system becomes operational, the master files must be created. This can be a major task, and it may involve the production of a file-conversion system, with its own programs.
Education and Training
An essential feature of the implementation of a new computer system is the education and training of all staff associated with it. The team has to educate the users about the project, though this activity would have been done during the systems analysis phase, it is worth being carried out here too. All staff must appreciate the objectives of the new system, and how it will operate, as well as the facilities it will provide. Indeed training should be organized for all the levels of staff; management, technical staff and typical end-users.
The detailed training must be supported by adequate documentation. This will be the user manual which is written by the system designer. There must be practice in the training programme. Special practice sessions must be arranged, especially where there is a direct changeover, when there is no practice period of pilot or parallel running.
Changeover is the process of changing from the old system to the new system. The four main changeover strategies that are used are the parallel strategy, direct cutover strategy, phased approach strategy and pilot strategy. A perfectly sound system can be destroyed by poor changeover. To be successful, remember, changeover has to have the support and involvement of managers and the co-operation of systems staff and users. Thus, prior to changeover, management must verify that the system does actually satisfy defined information needs; that the equipment, software and staff necessary for successful changeover are available; that control and audit procedures are in existence to ensure system integrity; and that performance requirements have been established for the system’s assessment in operation.
Direct changeover denotes the situation where at a specified time the old system is switched off and the new switched on, that is the new system entirely replaces the old system at an agreed date. With this approach, the new system must have been thoroughly tested so as to minimise risks in the initial operation. There are several situations where the technique is applicable or unavoidable
- In very small systems it is often not worthwhile considering any other technique, owing to the inherent simplicity of the system.
- In very large systems it is sometimes not feasible to maintain two systems simultaneously (as in parallel and pilot running) owing to the work involved.
- Where there is little similarity between the old and new systems, the simultaneous running of both systems may be unhelpful.
- The attractiveness of this approach is that it is economical. This is because resources are saved as the method involves the immediate discontinuance of the old system.
- Fast to implement a new system.
- This approach is risky because in the event of problems or failure with the new system, there will be no other functional system to fall on.
- Staff have to be trained very well on the new system.
- It is difficult to be planned
In parallel changeover the old and new systems are run with the same data until there is confidence in the new system, whereupon the old system is dropped. Parallel changeover or parallel running of the old and the new systems simultaneously allows a comparison of output to be made between them. Any shortcomings of the new system can be rectified, and continuous cross-checks made. Proper timelines must be established and adhered to if this approach is used.
Parallel changeover is often used so that the old system may still be operated when there is a breakdown in the new system. What has to be remembered in this particular context is that running in parallel means double the cost. Another problem concerns the staff and other resources used to run the two systems together. There may well need to be separate controls for the two systems, to be maintained and then reconciled. The objective should be to terminate the running of the old system as soon as is conveniently possible.
- This is a safer method as the old system is kept operational while the new system is brought in.
- Much greater security.
- It allows results to be compared to ensure that the new system is working without any errors.
- This approach facilitates the training of staff and helps them to gain confidence in the new system.
- The cost of implementation is very expensive because of the need to operate the two systems at the same time
- Parallel running implementation also requires a lot of time and needs frequent maintenance.
- Greater effort is needed to run the two systems and to compare the outputs.
- It may not be very easy to revert to the old system should things go wrong. The new system may handle the information differently, making it awkward to compare outputs.
- Knowing when to make the actual changeover is not an easy decision.
Phased changeover is where the new system is introduced in phases or stages as each stage is implemented and operating correctly. The phases continue until the whole system becomes operational. The method thus consists of a series of direct changes. This method is used for very large information systems which possess many complex components and which cross organisational frontiers. The implementation of each phase can be controlled, and risk to the user department is thus reduced considerably.
This method allows easier transfer of staff and is probably the most satisfactory method of working, where it is possible. It permits thorough testing under real conditions while limiting the risk of system failure. It requires, however, that part of the system functioning can be conveniently separated from the rest. It also requires some additional staff effort in handling two different systems simultaneously.
- There is considerable control as only manageable parts of the program are being changed over at a time.
- Training can be completed in small parts
- A failure of the new system has minimal impact because it is only one small part.
- This implementation method takes more time to get the new system into operation.
- The system may not easily be split into phases.
- The phases may well be different in the two systems.
- The interfaces between remaining old system phases and the new system phases already changed over, are extremely difficult to control.
With this type of changeover, the new system is installed or experimented with a selected department or departments (pilot departments) of the organization. When the software has been used satisfactorily for a considerable time, then the software is deployed in the whole organization. By restricting the implementation to a selected department or a branch of the organisation, the risk of system failure is restricted to just the pilot site which will be handy to resolve.
- Failure or problems can be identified and addressed without wide-spread impact to the organization
- Considerable control is retained and no risks are taken.
- Time is needed to collect and collate the data to be used.
- The two systems may handle the data differently, making comparison of outputs difficult.
Commissioning of a New System
Commissioning of a new system is a key activity during the implementation phase. Commissioning is the process of assuring that all systems (hardware, software etc.) are designed, installed, tested, operated, and maintained according to the users’ requirements established at the beginning of the systems development process. The final process of the commissioning is to deploy the system. In a typical information system project, the project manager or the systems analyst is responsible for developing the commissioning and the handover plan, also to ensure that all commissioning and handover requirements are addressed and relevant stakeholders consulted as part of the process.