In the last article we prioritised the critical legacy systems that need action. Armed with your priority list it's now time to start planning what you are going to do.  In broad terms you have the following choices:

Re-write the system in a new technology

If the legacy system is a core function that gives your organisation a competitive advantage then a re-write isn't as crazy as it sounds. 

On the plus-side:

• The application is easier to support and maintain
• You can iron out some of the less desirable features
• If re-built properly then you can take advantage of the latest technology

This can be an expensive route unless the scope is kept under tight control.  It is all too easy to keep adding new features and before you know it several years have passed and you still don't have the replacement system.

Migrate to a packaged application

Depending on the legacy system there could be a packaged application that meets your needs.  You will need to invest some time documenting business requirements from the legacy system.  Using these you can scan the marketplace looking for systems that meet your requirements.

On the plus-side:

• The legacy application is removed completely
• The package vendor will provide support and maintenance taking another headache away
• You will benefit from future upgrades to the package

It is unlikely you will get an exact match on requirements so compromises will need to be made.  Although many vendors offer to customise their products be wary of this as you may find future upgrades problematic and expensive. 

Maintain the existing system for the foreseeable future

If you choose this route then the real decision is how to continue with maintenance.  You can use in-house staff for this task but is this best use of these resources?  Whilst some technical people like to do this work, many don't so you are at risk of them leaving.  Also wouldn't you want the in-house team to be building new systems as this is a better business investment?

Outsourcing maintenance is a viable option as there are vendors who can provide this service.  When going down this route it is critical to select a partner that you can have a good relationship with.  They could be supporting your application for many years to come so it is critical they work with you all the way. 

Decommission the system

If you find yourself in this position then count yourself lucky.  However, don't rush headlong into decommission without some basic checks:

• Are you certain nothing else depends on the system?  Go back to your systems map and check again.
• Do you have a restore plan?  Despite all your checks there might still be something you missed so can you get the system back if you absolutely need to.
• Have you communicated to everyone?  Not just communicated but received their buy-in.  There are plenty of senior managers in organisations who love their pet systems and get pretty unhappy when they are turned off.

No matter which route you decide to take someone will need to approve the budget so make sure you have a good business case.