Information Technology Solutions
A careful well-devised plan for the implementation of a new system is important to ensure a smooth change over from the old to the new. Other reasons for advanced planning include:
The need to have a fallback position in case the system fails to operate successfully. The major activities during the installation of the new system include:
Parallel running means processing current data simultaneously by both old and new systems in order to check the results.
The current system is not dispensed with until the integrity of the new system has been proved. This method lends itself where the old and the new system have similar outputs, which can be cross-checked.
This changeover approach is generally suitable where the new system is not very labour intensive so that operating both systems does not over tax employee resources. Parallel running involves starting the whole system at the one time and therefore, if the system is not relatively, difficulties may be encountered - thus this from of changeover is more suited for small or medium systems.
As this is a fail-safe method it is costly. It is necessary to engage staff for running both systems side by side and moreover, it will prolong the testing period and delay changeover. Limits must therefore be set on the parallel cycles.
The failure rate is generally low when this method of changeover is used.
This is the process of progressively cutting small segments of the old system across to the new system. This may be done by periods of parallel running for each segment before cutting across or, alternatively, cutting a segment across without running in parallel first. The benefits of a gradual changeover are that the burden of the changeover is reduced and spread over a longer period of time, and there is an opportunity to learn from the previous changeover and its problems before the next one takes place.
This changeover approach is most suitable for large projects and for projects involving a large number of personnel, when it proves to impracticable to consider operating the old and new systems in parallel.
This method lends itself for installing a system where there are natural segments in the user population such as is provided by the structure of an organisation, for example, where a new payroll system for a business is based on a divisional structure, it may be decided to gradually introduced the system on a new divisional basis.
The expection is that the failure rate to be medium and costs to be in the medium range - less than for the parallel method.
It is claimed that pilot changeover is strictly an extended testing situation. The concept here is that data from a previous cycle of the old system is taken on the new system and the go-ahead to change to the new system is dependent on the approval of results of the pilot run. In practise, this is a large-scale system test being carried out as a preliminary to an immediate changeover.
This method is often used if the new system is to be installed in a number of different locations - at branches, depots, shops etc.
Normally one typical location is chosen for the pilot run. This allows the design team the chance to concentrate their efforts in the one area until the system has been shown to be workable.
This means that the system will be introduced when the results of this test meet the criteria expected.
The costs of this method of changeover would be medium and it would be expected that the failure rate of the new system would be low.
This method is also referred to as direct changeover.
Immediate is the name given to the method which leads to the complete replacement of the old system by the new one at a point in time.
The old system operates until the end of a week or a month, say, and at the beginning of the next week or month, the new system starts.
There is no fall back position if the new system should be found to have unforeseen errors or problems after a few weeks or months, because the old system has not been in operation over that period.
Thus, whilst immediate changeover is the simplest and least expensive method, it is also the most risky.
It requires the utmost confidence in the new system on the part of both users and computer staff and is best employed when the users have some previous experience of computerisation than when they are complete novices.
It tends to be used when the old system is not directly comparable with the old one - no benefit is gained by continuing to operate the old system - or when the timescale of change over is very tight, or when resources prevent a more thorough checking exercise such as is provided by parallel running.
It is imperative with an immediate changeover that prior testing of the system has been exhaustive.
This method is relatively cheap in terms of costs but the failure rate can be exceedingly high if the whole project has not been thoroughly tested planned
