The process of archiving data can be expensive, from the movement and testing of data, the creation of reports and the documentation of the process. It is necessary when the data is subject to records retention and needs to preserved for a long period.
Another option is to retain the application until the retention period is completed. I’ve written about the risks of retaining applications too long, but there are circumstances where this does make financial sense.
Til Death Do You Part
If the data retention period for data from a legacy application is short (more on that in a moment) then the option of keeping the application running might be viable. Reduce the operating costs as much as possible, limit access to only key business and IT users and run the clock out. When the retention period is over the application can be retired and the data purged without archiving.
There is no single answer on whether to run out the clock or to perform an archive, but the same criteria can be applied to make the decision.
Balancing Risks and Costs
As mentioned in another post retaining technical debt incurs both cost and risk. Although there can be significant savings achieved by planning the retirement correctly there will still be multiple environments for every application that need to be supported, patches have to be applied, software licenses paid, maintenance contracts may be required. For older applications where software, hardware or operating systems are no longer supported the cyber risk will increase, and for older hardware the risk of failure increases.
It is important to balance the increasing technical debt of the aging application against the cost to perform an archive, and to be realistic about the costs. Also factor in the possibility of the duration that the application has to be sustained may be extended by a legal hold or other factors.
Define “Short”
Running out the clock can be a cost effective if the retention period is “short”, which of course is a highly subjective consideration. More than 2 years is almost certainly too long, 1-2 years would depend on the cost to archive.
Temporary Stay of Execution
A variant on running out the clock is where there are complex reporting requirements that are only needed for a limited period. An accounts payable application is an example of this where the reporting requirements are significant for about 18-24 months after the end of the financial year, and then up to 80% of reports are no longer required. In these cases it can be cost effective to delay the retirement of the application until the reporting requirements diminish which will reduce the cost of performing the archive.
In this case it might still be appropriate to archive the data immediately and develop the limited set of reports that will be required in the long term future. This can be driven by the availability of people knowledgeable in the legacy application, the availability of project funds to complete the archive or even just because applications left on the shelf tend to stay longer.
A temporary stay of execution should only be granted with a clear business justification and with a solid plan for when and how the application retirement will be completed. These temporary stays have a tendency to become semi-permanent …
Conclusion
Keeping legacy applications running for a period of time can be appropriate and aligned with good data governance. However the practice should be treated with caution as the cost/benefit of doing so will change if the proposed limited period for keeping the application running starts to be extended.
